Account checking method and device, computer device, storage medium and program product

Abstract

This application relates to a reconciliation method, apparatus, computer equipment, storage medium, and program product. The method includes: acquiring real-time job data from multiple computing clusters; aggregating the real-time job data from the multiple computing clusters according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics; and reconciling the real-time job data from the multiple computing clusters based on the aggregation results corresponding to the different aggregation metrics to generate a reconciliation result. This method can perform reconciliation for multiple computing clusters with high complexity and large data volumes.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a reconciliation method, apparatus, computer equipment, storage medium, and program product. Background Technology

[0002] When accounts execute jobs across multiple high-performance computing clusters, each job consumes certain computing resources (such as CPU and GPU), and the price of these resources varies for each job. Therefore, it is necessary to record and reconcile the expenses incurred by each account during job execution. Generally, this can be done through a cloud platform.

[0003] However, traditional reconciliation methods on cloud platforms can only reconcile accounts for single computing clusters with identical resource types, prices, and computing power. When multiple computing clusters are connected to a cloud platform, these clusters contain different resource types, computing power levels, and corresponding resource prices. Therefore, traditional reconciliation methods for single computing clusters are inadequate for reconciling accounts for highly complex computing clusters with large data volumes. Summary of the Invention

[0004] Therefore, it is necessary to provide a reconciliation method, apparatus, computer equipment, storage medium, and program product that can reconcile accounts for multiple computing clusters with high complexity and large data volume, in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides a reconciliation method. The method includes:

[0006] Acquire real-time job data from multiple computing clusters;

[0007] The real-time job data in the multiple computing clusters are aggregated according to different aggregation metrics to generate aggregation results corresponding to the different aggregation metrics.

[0008] Based on the aggregation results corresponding to the different aggregation metrics, the real-time job data in the multiple computing clusters are reconciled to generate reconciliation results.

[0009] This application can aggregate real-time job data from multiple computing clusters according to different aggregation metrics. This allows for the classification and statistical analysis of highly complex, large-volume real-time job data from multiple computing clusters based on these metrics, thereby generating relatively accurate aggregation results for different aggregation metrics across multiple computing clusters. Subsequently, based on these accurate aggregation results for different aggregation metrics across multiple computing clusters, reconciliation of the highly complex, large-volume real-time job data from multiple computing clusters can be performed relatively accurately, resulting in a more precise reconciliation result. Therefore, this application can reconcile real-time job data from multiple highly complex, large-volume computing clusters based on different aggregation metrics.

[0010] In one embodiment, the different aggregation metrics include a first type of aggregation metrics and a second type of aggregation metrics; the aggregation of real-time job data in the multiple computing clusters according to the different aggregation metrics to generate aggregation results corresponding to the different aggregation metrics includes:

[0011] The real-time job data in the multiple computing clusters are aggregated according to the first type of aggregation index to generate a first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes a billing index and the total resource consumption corresponding to the billing index; the billing index includes a cluster identifier, a user identifier, a resource type, and a resource unit price.

[0012] The real-time job data in the multiple computing clusters are aggregated according to the second type of aggregation index to generate a second aggregation result corresponding to the second type of aggregation index; the second type of aggregation index includes the billing index, the resource consumption of each job executed under the billing index, and the job cost data of each job executed under the billing index.

[0013] In this embodiment, real-time job data from multiple computing clusters with high complexity and large data volume can be aggregated according to different cluster identifiers, user identifiers, resource types, resource unit prices, and total resource consumption corresponding to different billing indicators. This allows for the accurate generation of the first aggregation result corresponding to the first type of aggregation indicator. Alternatively, real-time job data from multiple computing clusters with high complexity and large data volume can be aggregated according to different cluster identifiers, user identifiers, resource types, resource unit prices, resource consumption for each job executed under the billing indicator, and job cost data for each job executed under the billing indicator. This simplifies the pipeline data corresponding to user job data and allows for the accurate generation of the second aggregation result corresponding to the second type of aggregation indicator.

[0014] In one embodiment, the step of reconciling the real-time job data in the multiple computing clusters based on the aggregation results corresponding to the different aggregation metrics, and generating reconciliation results, includes:

[0015] The total resource consumption corresponding to the same resource type and the same resource unit price is determined from the first aggregation result. Based on the total resource consumption and the resource unit price, the total first operation cost corresponding to the same resource type and the same resource unit price is calculated.

[0016] From the second aggregation result, determine the operation cost data for each operation under the same resource type and the same resource unit price, and determine the total second operation cost based on the operation cost data for each operation;

[0017] Based on the total cost of the first job and the total cost of the second job, the real-time job data in the multiple computing clusters are reconciled to generate the reconciliation result.

[0018] In this embodiment, for the same resource type and unit price, the total resource consumption corresponding to the same resource type and unit price can be accurately determined from the first aggregation result. Therefore, the total first operation cost can be calculated more accurately based on this accurate total resource consumption and unit price. Furthermore, the operation cost data for each operation under the same resource type and unit price can be accurately determined from the second aggregation result. Therefore, the total second operation cost can be accurately determined based on the operation cost data for each operation. Then, based on the accurate total first operation cost and the accurate total second operation cost, real-time operation data in multiple computing clusters with high complexity and large data volumes can be reconciled more accurately, thereby generating a more accurate reconciliation result.

[0019] In one embodiment, the step of reconciling the real-time job data in the plurality of computing clusters based on the first total job cost and the second total job cost, and generating the reconciliation result, includes:

[0020] If the total cost of the first operation is equal to the total cost of the second operation, then the reconciliation result is determined to be a successful reconciliation.

[0021] If the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result;

[0022] The total first resource consumption and the total second resource consumption corresponding to each of the aforementioned billing indicators are compared to generate the reconciliation result.

[0023] In this embodiment, if the total cost of the first job is equal to the total cost of the second job, the reconciliation result can be determined as a successful reconciliation. If the total cost of the first job is not equal to the total cost of the second job, the total resource consumption in the first aggregation result and the total resource consumption in the second aggregation result are reconciled from the perspective of the total resource consumption corresponding to each billing indicator. This can determine the range of real-time job data reconciliation failures in multiple computing clusters, thereby generating more accurate reconciliation results and reducing the difficulty of subsequent investigation into the location of reconciliation failures.

[0024] In one embodiment, acquiring real-time job data from multiple computing clusters includes:

[0025] Within a preset time period, the job data of each job executed by the multiple computing clusters under the billing indicators is obtained, and based on the job data of each job, the first job data total of all jobs executed by the multiple computing clusters within the preset time period is calculated.

[0026] According to at least one of the billing metrics, the total amount of second job data for all jobs executed by the plurality of computing clusters within the preset time period is obtained; the total amount of second job data includes the sum of the resource consumption for each job executed under the billing metrics;

[0027] Based on the total amount of the first job data and the total amount of the second job data, the real-time job data in the plurality of computing clusters is determined.

[0028] In this embodiment, within a preset time period, job data for each job executed by multiple computing clusters under the billing indicators can be obtained relatively accurately from multiple computing clusters according to the billing indicators. Based on the job data for each job, the total first job data volume for all jobs executed by multiple computing clusters within the preset time period can be calculated relatively accurately. Furthermore, a second job data volume for all jobs executed by multiple computing clusters within the preset time period can be obtained relatively accurately from multiple computing clusters according to at least one of the billing indicators. Then, based on the relatively accurate first job data volume and the relatively accurate second job data volume, the real-time job data in multiple computing clusters with high complexity and large data volume can be determined relatively accurately.

[0029] In one embodiment, determining the real-time job data in the plurality of computing clusters based on the total amount of the first job data and the total amount of the second job data includes:

[0030] Compare the total amount of the first task data with the total amount of the second task data to generate a comparison result;

[0031] If the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data, then the total amount of the first job data or the total amount of the second job data shall be used as the real-time job data in the plurality of computing clusters.

[0032] In this embodiment, the total amount of the first job data is compared with the total amount of the second job data to generate a comparison result. If the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data, it indicates that the total amount of job data transmitted from multiple computing clusters to the cloud platform is accurate. Therefore, either the total amount of the first job data or the total amount of the second job data can be used as the real-time job data in multiple computing clusters, thus determining relatively accurate real-time job data in multiple computing clusters with high complexity and large data volumes. Therefore, this embodiment can obtain job data for each group of users in each computing cluster for different resource types based on the dimension of multiple computing clusters, and periodically compare it with the detailed job data obtained from the cloud platform, thereby ensuring the consistency between the job data of multiple computing clusters and the cloud platform.

[0033] In one embodiment, the method further includes:

[0034] If the comparison result is that the total amount of the first job data is not equal to the total amount of the second job data, then within the preset time period, the job data of each job executed by the multiple computing clusters is re-acquired, and based on the job data of each job, the new total amount of the first job data of all jobs executed by the multiple computing clusters within the preset time period is calculated.

[0035] According to at least one of the billing metrics, the total amount of new second job data for all jobs executed by the multiple computing clusters is reacquired within the preset time period;

[0036] Based on the total amount of the new first job data and the total amount of the new second job data, new real-time job data is determined in the plurality of computing clusters.

[0037] In this embodiment, if the comparison result shows that the total amount of the first job data is not equal to the total amount of the second job data, it indicates that an error occurred during the transmission of the total amount of the first job data and / or the total amount of the second job data during the transmission of the first job data and / or the second job data during the transmission of the first job data and / or the second job data during the transmission of the first job data and the second job data during the transmission of the second job data within the transmission of the second job data. Therefore, the new real-time job data in the multiple computing clusters can be re-determined based on the re-acquired total amount of the first job data and the re-acquired total amount of the second job data. Thus, this embodiment proposes a method for data processing in cases where job data is abnormal (i.e., inconsistencies in job data occur during reconciliation).

[0038] Secondly, this application also provides a reconciliation device. The device includes:

[0039] The acquisition module is used to acquire real-time job data from multiple computing clusters;

[0040] The aggregation module is used to aggregate real-time job data in the multiple computing clusters according to different aggregation metrics, and generate aggregation results corresponding to the different aggregation metrics.

[0041] The reconciliation module is used to reconcile the real-time job data in the multiple computing clusters based on the aggregation results corresponding to the different aggregation indicators, and generate reconciliation results.

[0042] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method in any of the embodiments of the first aspect described above.

[0043] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.

[0044] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.

[0045] The aforementioned reconciliation method, apparatus, computer equipment, storage medium, and program product acquire real-time job data from multiple computing clusters; aggregate the real-time job data from multiple computing clusters according to different aggregation metrics, generating aggregation results corresponding to different aggregation metrics; and reconcile the real-time job data from multiple computing clusters based on the aggregation results corresponding to different aggregation metrics, generating reconciliation results. This application can aggregate the acquired real-time job data from multiple computing clusters according to different aggregation metrics, enabling the classification and statistical analysis of real-time job data from multiple computing clusters with high complexity and large data volumes according to different aggregation metrics, thereby generating relatively accurate aggregation results corresponding to different aggregation metrics for multiple computing clusters. Subsequently, based on the relatively accurate aggregation results corresponding to different aggregation metrics for multiple computing clusters, the real-time job data from multiple computing clusters with high complexity and large data volumes can be reconciled relatively accurately, thus generating relatively accurate reconciliation results. Therefore, this application can reconcile real-time job data from multiple computing clusters with high complexity and large data volumes based on different aggregation metrics. Attached Figure Description

[0046] Figure 1 This is a diagram illustrating the application environment of the reconciliation method in one embodiment;

[0047] Figure 2 This is a flowchart illustrating the reconciliation method in one embodiment;

[0048] Figure 3 This is a flowchart illustrating the aggregation steps in one embodiment;

[0049] Figure 4 This is a schematic diagram of the process for generating the first aggregation result and the second aggregation result in one embodiment;

[0050] Figure 5 This is a flowchart illustrating the reconciliation steps in one embodiment;

[0051] Figure 6 This is a schematic diagram of the reconciliation process in one embodiment;

[0052] Figure 7 This is a flowchart illustrating the steps for generating reconciliation results in one embodiment;

[0053] Figure 8 This is a flowchart illustrating the real-time job data determination steps in one embodiment;

[0054] Figure 9 This is a schematic diagram of the structure in one embodiment where multiple computing clusters transmit job data to a cloud platform.

[0055] Figure 10This is a flowchart illustrating the steps for determining real-time job data where the comparison results are equal in one embodiment.

[0056] Figure 11 This is a flowchart illustrating the processing of abnormal job data in one embodiment;

[0057] Figure 12 This is a flowchart illustrating the steps for determining real-time job data where the comparison results are unequal, as shown in another embodiment.

[0058] Figure 13 This is a flowchart illustrating the reconciliation method in one optional embodiment;

[0059] Figure 14 This is a schematic diagram illustrating the overall process of reconciling real-time job data across multiple clusters in one embodiment.

[0060] Figure 15 This is a structural block diagram of the reconciliation device in one embodiment;

[0061] Figure 16 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0063] When accounts execute jobs across multiple high-performance computing clusters, each job consumes certain computing resources (such as CPU and GPU), and the price of these resources varies for each job. Therefore, it is necessary to record and reconcile the expenses incurred by each account during job execution. Generally, this can be done through a cloud platform.

[0064] However, traditional reconciliation methods on cloud platforms can only reconcile accounts for single computing clusters with identical resource types, prices, and computing power. When multiple computing clusters are connected to a cloud platform, these clusters contain different resource types, computing power levels, and corresponding resource prices. Therefore, traditional reconciliation methods for single computing clusters are inadequate for reconciling accounts for highly complex computing clusters with large data volumes.

[0065] The reconciliation method provided in this application embodiment can be applied to, for example, Figure 1In the application environment shown, cloud platform 102 communicates with computing cluster 104 via a network. A data storage system can store the data that computing cluster 104 needs to process. The data storage system can be integrated onto computing cluster 104 or placed in the cloud or on other network servers. Cloud platform 102 acquires real-time job data from multiple computing clusters; cloud platform 102 aggregates the real-time job data from multiple computing clusters according to different aggregation metrics, generating aggregation results corresponding to different aggregation metrics; cloud platform 102 reconciles the real-time job data from multiple computing clusters based on the aggregation results corresponding to different aggregation metrics, generating reconciliation results. Cloud platform 102 can be a server. Computing cluster 104 can be implemented using a standalone server or a server cluster composed of multiple servers.

[0066] In one embodiment, such as Figure 2 As shown, a reconciliation method is provided, which can be applied to... Figure 1 Taking cloud platform 102 as an example, the explanation includes the following steps:

[0067] S220 acquires real-time job data from multiple computing clusters.

[0068] In this context, a cloud platform, also known as a cloud computing platform, is an operational platform that processes computing resources to create various services. Multiple computing clusters can be high-performance computing (HPC) clusters. A job refers to a task submitted by a user that needs to be executed within the computing cluster. Real-time job data refers to data related to a job generated during its execution.

[0069] Optionally, since each user can execute jobs through multiple computing clusters, real-time job data from multiple computing clusters will be generated during the execution of jobs through multiple computing clusters. Optionally, the cloud platform 102 can directly obtain the real-time job data from multiple computing clusters; or, the cloud platform 102 can first obtain job data within a preset time period from multiple computing clusters, and then determine the real-time job data from multiple computing clusters within the preset time period based on the job data within the preset time period. The preset time period can be set according to the actual job execution status, and this embodiment does not limit this.

[0070] S240 aggregates real-time job data from multiple computing clusters according to different aggregation metrics, generating aggregation results corresponding to different aggregation metrics.

[0071] Optionally, since the resource types, unit prices, and other data used in executing different jobs vary, the cloud platform 102 can set different aggregation metrics for each real-time job data. These aggregation metrics are the aggregation dimensions used in the process of aggregating real-time job data, and may include, but are not limited to, cluster identifier, user identifier, resource type, resource unit price, resource consumption, and job data volume. Job data volume represents the number of jobs executed by the user.

[0072] Subsequently, cloud platform 102 can aggregate real-time job data from multiple computing clusters according to different aggregation metrics, generating aggregation results corresponding to different aggregation metrics. For example, assuming the aggregation metrics are resource type, resource unit price, and job data volume, cloud platform 102 can classify and statistically analyze the real-time job data from multiple computing clusters according to different resource types, different resource unit prices, and different job data volumes, generating aggregation results corresponding to different resource types, different resource unit prices, and different job data volumes. Data aggregation refers to merging data from different data sources for the same object to obtain more complete information about that object. Aggregation results corresponding to different aggregation metrics refer to the statistical results of real-time job data obtained by aggregating real-time job data from multiple computing clusters according to different aggregation metrics.

[0073] S260 reconciles real-time job data across multiple computing clusters based on the aggregation results corresponding to different aggregation metrics, and generates reconciliation results.

[0074] Optionally, the cloud platform 102 can directly reconcile real-time job data across multiple computing clusters based on the aggregation results corresponding to different aggregation metrics, generating reconciliation results. Alternatively, the cloud platform 102 can first calculate job cost data corresponding to different aggregation metrics based on the aggregation results, and then reconcile real-time job data across multiple computing clusters based on the job cost data corresponding to different aggregation metrics, generating reconciliation results. Reconciliation refers to verifying the accuracy of cost data in the accounts. The reconciliation result can include successful or unsuccessful reconciliation.

[0075] The aforementioned reconciliation method involves acquiring real-time job data from multiple computing clusters; aggregating the real-time job data from multiple computing clusters according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics; and reconciling the real-time job data from multiple computing clusters based on the aggregation results corresponding to different aggregation metrics to generate reconciliation results. This application can aggregate the acquired real-time job data from multiple computing clusters according to different aggregation metrics, enabling the classification and statistical analysis of real-time job data from multiple computing clusters with high complexity and large data volumes according to different aggregation metrics. This allows for the generation of more accurate aggregation results corresponding to different aggregation metrics for multiple computing clusters. Subsequently, based on the more accurate aggregation results corresponding to different aggregation metrics for multiple computing clusters, the real-time job data from multiple computing clusters with high complexity and large data volumes can be reconciled more accurately, thus generating more accurate reconciliation results. Therefore, this application can reconcile real-time job data from multiple computing clusters with high complexity and large data volumes based on different aggregation metrics.

[0076] The above embodiments involve aggregating real-time job data from multiple computing clusters according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics. The specific methods are described below. In one embodiment, the different aggregation metrics include a first type of aggregation metric and a second type of aggregation metric, such as... Figure 3 As shown, S240 includes:

[0077] S320 aggregates real-time job data from multiple computing clusters according to the first type of aggregation index, generating the first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes billing index and the total resource consumption corresponding to the billing index; the billing index includes cluster identifier, user identifier, resource type and resource unit price.

[0078] The various aggregated metrics can include, but are not limited to, billing metrics, the total resource consumption corresponding to the billing metrics, the amount of job data, the resource consumption for each job executed under the billing metrics, and the job cost data for each job executed under the billing metrics. Billing metrics include cluster identifier, user identifier, resource type, and resource unit price. The cluster identifier represents different computing clusters. The user identifier represents different users or different accounts. Generally, a research team uses the same account for billing and reconciliation. The resource type refers to the model of computing resources used by each user in multiple computing clusters when executing jobs. Computing resources can include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), and storage resources. The resource unit price refers to the unit price of each computing resource. For example, assuming the resource type of the computing resource is model A, the resource unit price of this computing resource is XX yuan / core.

[0079] In this context, "job data volume" refers to the number of jobs executed under the same billing metric. "Resource consumption," also known as "machine time," refers to the amount of computing resources consumed during job execution, measured in core hours or card hours. For example, if a job runs for one hour and uses two CPU cores, its resource consumption or machine time is 2 core hours. The total resource consumption corresponding to a billing metric is the sum of all resource consumption values ​​based on the billing metric. "Job cost data" refers to the cost incurred for each job executed under the billing metric. For example, the job cost data for executing job A is equal to the unit price of the computing resources used in executing job A multiplied by the resource consumption of the computing resources used in executing job A. In this embodiment, different aggregated metrics can be divided into a first type of aggregated metrics and a second type of aggregated metrics. The first type of aggregated metrics includes, but is not limited to, billing metrics and the total resource consumption corresponding to the billing metrics.

[0080] Optionally, since different aggregation metrics include the first type of aggregation metrics, the cloud platform 102 can aggregate real-time job data from multiple computing clusters according to the first type of aggregation metrics to generate a first aggregation result corresponding to the first type of aggregation metrics. That is, the cloud platform 102 can classify and statistically analyze real-time job data from multiple computing clusters according to different cluster identifiers, different user identifiers, different resource types, different resource unit prices, and different total resource consumption corresponding to different billing metrics, thereby generating a first aggregation result corresponding to different cluster identifiers, different user identifiers, different resource types, different resource unit prices, and different total resource consumption. The first aggregation result refers to the statistical result of the real-time job data obtained by aggregating real-time job data from multiple computing clusters according to the first type of aggregation metrics.

[0081] like Figure 4 As shown, Figure 4 This is a flowchart illustrating the generation of a first aggregation result and a second aggregation result in one embodiment. First, in step S402, the cloud platform 102 parses real-time job data from multiple computing clusters to obtain job data corresponding to each job, including cluster identifiers, user identifiers, resource types, job durations, resource unit prices, and resource consumption. Second, in step S404, the cloud platform 102 calculates the job data volume corresponding to each billing indicator and the total resource consumption corresponding to each billing indicator based on the aforementioned job data. Then, in step S406, the cloud platform 102 aggregates the real-time job data from multiple computing clusters based on the aforementioned job data, the job data volume corresponding to each billing indicator, and the total resource consumption corresponding to each billing indicator, thereby generating a first aggregation result. The first aggregation result can be in the form of a job aggregation table. For example, a job aggregation table for a specific computing cluster can be shown in the following table.

[0082] Table 1

[0083] Aggregate table identifier User ID Resource types resource unit price Job data volume Total resource consumption 10000001 Account A Model A 2.0 yuan / core time 100 500 cores 10000002 Account A Model A 1.8 yuan / core hour 200 320 cores 10000003 Account A Model B 0.2 yuan / core time 150 200 cores 10000004 Account A Model C 0.5 yuan / core time 1000 3000 cores 10000005 Account B Model A 2.2 yuan / core time 600 1000 cores 10000006 Account D Model D 1.2 yuan / core hour 200 500 cores

[0084] S340, aggregate real-time job data from multiple computing clusters according to the second type of aggregation index to generate the second aggregation result corresponding to the second type of aggregation index; the second type of aggregation index includes billing index, resource consumption for each job executed under the billing index, and job cost data for each job executed under the billing index.

[0085] Optionally, since different aggregation metrics include second-type aggregation metrics, and the second-type aggregation metrics include, but are not limited to, billing metrics, resource consumption for each job executed under the billing metrics, and job cost data for each job executed under the billing metrics, the cloud platform 102 can aggregate real-time job data from multiple computing clusters according to the second-type aggregation metrics to generate a second aggregation result corresponding to the second-type aggregation metrics. That is, the cloud platform 102 can classify and statistically analyze real-time job data from multiple computing clusters according to different cluster identifiers, different user identifiers, different resource types, different resource unit prices, resource consumption for each job executed under the billing metrics, and job cost data for each job executed under the billing metrics, thereby generating a second aggregation result corresponding to different cluster identifiers, different user identifiers, different resource types, different resource unit prices, resource consumption for each job executed under the billing metrics, and job cost data for each job executed under the billing metrics. The second aggregation result refers to the statistical result of the real-time job data obtained by aggregating real-time job data from multiple computing clusters according to the second-type aggregation metrics.

[0086] Combination Figure 4 As shown, firstly, in step S402, the cloud platform 102 can parse the real-time job data from multiple computing clusters to obtain job data such as cluster identifiers, user identifiers, resource types, job durations, resource unit prices, and resource consumption corresponding to each job. Secondly, in step S404, the cloud platform 102 can calculate the resource consumption and job cost data for each job executed under the billing indicators based on the above job data. Then, in step S408, the cloud platform 102 can aggregate the real-time job data from multiple computing clusters based on the above job data, the resource consumption and job cost data for each job executed under the billing indicators, thereby generating a second aggregation result. The second aggregation result can be in the form of an account transaction table. For the account transaction table, each time job cost data is recorded, the aggregation table identifier corresponding to that job needs to be recorded synchronously. For example, the account transaction table for a certain computing cluster can be shown in the following table.

[0087] Table 2

[0088] Aggregate table identifier User ID Resource types resource unit price Operation cost data resource consumption 10000001 Account A Model A 2.0 yuan / core time 100 yuan 50 cores 10000001 Account A Model A 2.0 yuan / core time 250 yuan 125 cores 10000001 Account A Model A 2.0 yuan / core time 400 yuan 200 cores 10000002 Account A Model A 1.8 yuan / core hour 270 yuan 150 cores 10000003 Account A Model B 0.2 yuan / core time 20 yuan 100 cores 10000005 Account B Model A 2.2 yuan / core time 440 yuan 200 cores ...... ...... ...... ...... ...... ......

[0089] In this embodiment, real-time job data from multiple computing clusters with high complexity and large data volume can be aggregated according to different cluster identifiers, user identifiers, resource types, resource unit prices, and total resource consumption corresponding to different billing indicators. This allows for the accurate generation of the first aggregation result corresponding to the first type of aggregation indicator. Alternatively, real-time job data from multiple computing clusters with high complexity and large data volume can be aggregated according to different cluster identifiers, user identifiers, resource types, resource unit prices, resource consumption for each job executed under the billing indicator, and job cost data for each job executed under the billing indicator. This simplifies the pipeline data corresponding to user job data and allows for the accurate generation of the second aggregation result corresponding to the second type of aggregation indicator.

[0090] The above embodiments involve reconciling real-time job data in multiple computing clusters based on aggregation results corresponding to different aggregation metrics, and generating reconciliation results. The specific methods are described below. In one embodiment, such as... Figure 5 As shown, S260 includes:

[0091] S520, determine the total resource consumption corresponding to the same resource type and the same resource unit price from the first aggregation result, and calculate the total first operation cost corresponding to the same resource type and the same resource unit price based on the total resource consumption and the resource unit price.

[0092] Optionally, such as Figure 6 As shown, Figure 6 This is a schematic diagram of the reconciliation process in one embodiment. S602, the cloud platform 102 can set a reconciliation period and query the first aggregation result to determine the total resource consumption corresponding to the same resource type and unit price within the reconciliation period. Then, S604, the cloud platform 102 can calculate the total first operating cost corresponding to the same resource type and unit price within the reconciliation period based on the total resource consumption corresponding to the same resource type and unit price within the reconciliation period, and the unit price corresponding to the same resource type and unit price within the reconciliation period. The total first operating cost corresponding to the same resource type and unit price within the reconciliation period is equal to the product of the total resource consumption corresponding to the same resource type and unit price within the reconciliation period and the unit price corresponding to the same resource type and unit price within the reconciliation period.

[0093] S540, determine the operation cost data for each operation under the same resource type and the same resource unit price from the second aggregation result, and determine the total second operation cost based on the operation cost data for each operation.

[0094] Optionally, combined Figure 6As shown in step S606, the cloud platform 102 can query the second aggregation result based on the set reconciliation period, and then determine the operation cost data for each operation under the same resource type and the same resource unit price within the reconciliation period from the second aggregation result. Then, in step S608, the cloud platform 102 can calculate the total second operation cost corresponding to the same resource type and the same resource unit price within the reconciliation period based on the operation cost data for each operation under the same resource type and the same resource unit price within the reconciliation period. The total second operation cost corresponding to the same resource type and the same resource unit price within the reconciliation period is equal to the sum of the operation cost data for each operation under the same resource type and the same resource unit price within the reconciliation period.

[0095] S560 reconciles real-time job data across multiple computing clusters based on the total cost of the first job and the total cost of the second job, and generates reconciliation results.

[0096] Optionally, for the same resource type and unit price within the reconciliation period, the cloud platform 102 can compare the total first operation cost with the total second operation cost to generate a comparison result of the total operation cost. This comparison can be made by subtracting the total first operation cost from the total second operation cost; or by finding the quotient of the total first operation cost from the total second operation cost; or by determining whether the total first operation cost is equal to the total second operation cost. Of course, this embodiment does not limit the method of comparing the total first operation cost with the total second operation cost. The comparison result of the total operation cost refers to the result obtained after comparing the total first operation cost with the total second operation cost. For example, assuming this embodiment compares the total first operation cost with the total second operation cost by determining whether they are equal, then the comparison result of the total operation cost is either that the total operation cost is equal or that the total operation cost is unequal. Afterwards, the cloud platform 102 can reconcile the real-time operation data in multiple computing clusters based on the comparison result of the total operation cost and generate a reconciliation result.

[0097] In this embodiment, for the same resource type and unit price, the total resource consumption corresponding to the same resource type and unit price can be accurately determined from the first aggregation result. Therefore, the total first operation cost can be calculated more accurately based on this accurate total resource consumption and unit price. Furthermore, the operation cost data for each operation under the same resource type and unit price can be accurately determined from the second aggregation result. Therefore, the total second operation cost can be accurately determined based on the operation cost data for each operation. Then, based on the accurate total first operation cost and the accurate total second operation cost, real-time operation data in multiple computing clusters with high complexity and large data volumes can be reconciled more accurately, thereby generating a more accurate reconciliation result.

[0098] The above embodiment involves reconciling real-time job data across multiple computing clusters based on the total first job cost and the total second job cost, generating reconciliation results. The specific method is described below. In one embodiment, such as... Figure 7 As shown, S560 includes:

[0099] S720, if the total cost of the first operation is equal to the total cost of the second operation, then the reconciliation result is determined to be successful.

[0100] Optionally, combined Figure 6 As shown in step S610, for the same resource type and unit price within the reconciliation period, the cloud platform 102 can detect whether the total cost of the first operation is equal to the total cost of the second operation. If the total cost of the first operation is found to be equal to the total cost of the second operation, then in step S612, the cloud platform 102 can determine that the reconciliation result is successful, record this reconciliation with a successful result, and update the user's account information and the time of successful reconciliation.

[0101] S740, if the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result.

[0102] Optionally, combined Figure 6As shown, if the total cost of the first job is not equal to the total cost of the second job, the cloud platform 102 can, for each billing indicator, S614, obtain the total first resource consumption corresponding to the billing indicator from the first aggregation result, and S616, obtain the total second resource consumption corresponding to the billing indicator from the second aggregation result. The total second resource consumption is the sum of the resource consumption of each job executed under the billing indicator in the second aggregation result. Referring to Tables 1 and 2, the aggregation table identifier is the billing indicator; that is, one billing indicator corresponds to one aggregation table identifier. For Table 1, each row represents the total resource consumption under each billing indicator; therefore, one row in Table 1 corresponds to one billing indicator. For Table 2, each row represents the job cost data and resource consumption for executing one job; therefore, multiple rows in Table 2 may correspond to one billing indicator.

[0103] For example, referring to Tables 1 and 2, for aggregate table identifier 10000001, that is, for the billing indicator (computing cluster A, account A, model A, 2.0 yuan / core time), the first total resource consumption (500 core time) corresponding to aggregate table identifier 10000001 can be obtained from Table 1; and the resource consumption of each job execution corresponding to aggregate table identifier 10000001 (50 core time, 125 core time, 200 core time) can be obtained from Table 2. The second total resource consumption (50 core time + 125 core time + 200 core time = 375 core time) can be obtained by adding the resource consumption of each job execution.

[0104] S760 compares the total first resource consumption and the total second resource consumption corresponding to each billing indicator and generates reconciliation results.

[0105] Optionally, the cloud platform 102 can compare the total first resource consumption and the total second resource consumption corresponding to each billing indicator to generate reconciliation results. Combined with... Figure 6As shown in step S618, the total first resource consumption and the total second resource consumption corresponding to each billing indicator are compared sequentially. If the total first resource consumption and the total second resource consumption corresponding to the same billing indicator are equal, then in step S620, for the next billing indicator, the total first resource consumption and the total second resource consumption corresponding to that next billing indicator are compared until the total first resource consumption and the total second resource consumption corresponding to each billing indicator have been compared, and the reconciliation ends. If the total first resource consumption and the total second resource consumption corresponding to the same billing indicator are not equal, then in step S622, the billing indicator and the corresponding operation data are recorded; then for the next billing indicator, the total first resource consumption and the total second resource consumption corresponding to that next billing indicator are compared until the total first resource consumption and the total second resource consumption corresponding to each billing indicator have been compared, and the reconciliation ends.

[0106] In this embodiment, if the total cost of the first job is equal to the total cost of the second job, the reconciliation result is considered successful. If the total cost of the first job is not equal to the total cost of the second job, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result. The total second resource consumption is the sum of the resource consumption of each job executed under the billing indicator in the second aggregation result. The total first resource consumption and the total second resource consumption corresponding to each billing indicator are compared to generate a reconciliation result. This allows for the reconciliation of real-time job data in multiple computing clusters where the total cost of the first job is not equal to the total cost of the second job, from the perspective of the total resource consumption corresponding to each billing indicator. This helps determine the range of reconciliation failures in multiple computing clusters, thereby generating more accurate reconciliation results and reducing the difficulty of subsequent investigation into the location of reconciliation failures.

[0107] The above embodiments involve acquiring real-time job data from multiple computing clusters. The specific methods are described below. In one embodiment, such as... Figure 8 As shown, S220 includes:

[0108] S820 acquires job data for each job executed by multiple computing clusters under billing metrics within a preset time period, and calculates the total first job data of all jobs executed by multiple computing clusters within the preset time period based on the job data of each job.

[0109] Optionally, such as Figure 9 As shown, Figure 9This is a schematic diagram illustrating the structure of multiple computing clusters transmitting job data to a cloud platform in one embodiment. The cloud platform 102 can periodically retrieve job data from multiple computing clusters under billing indicators, according to different computing clusters, different user identifiers, and different resource types, within a preset time period. Furthermore, the cloud platform 102 can calculate the total job data of all jobs executed by multiple computing clusters within the preset time period based on the job data of each job executed by multiple computing clusters under the billing indicators, and aggregate this total job data according to different computing clusters, different user identifiers, and different resource types to obtain the first total job data of all jobs executed by multiple computing clusters within the preset time period. The preset time period can be set according to the performance of the cloud platform and the computing clusters; specifically, the preset time period can be equal to the reconciliation cycle. Of course, this embodiment does not limit the preset time period. The first total job data refers to the total amount of data used by multiple computing clusters in the process of executing all jobs within the preset time period. The first total job data may include, but is not limited to, job volume, resource consumption, and storage volume. Job load refers to the number of jobs executed by multiple computing clusters within a preset time period, while storage capacity refers to the storage capacity used during job execution, such as 500GB or 1.2T.

[0110] S840, according to at least one of the billing indicators, within a preset time period, obtain the total amount of second job data for all jobs executed by multiple computing clusters; the total amount of second job data includes the sum of resource consumption for each job executed under the billing indicators.

[0111] Optionally, combined Figure 9 As shown, the cloud platform 102 can obtain detailed job data of all jobs executed by multiple computing clusters in real time from multiple computing clusters through message queues within a preset time period. The detailed job data may include, but is not limited to, job execution duration, resource consumption, and queues used during job execution. Furthermore, the cloud platform 102 can aggregate the detailed job data within the preset time period according to at least one of the billing metrics, thereby obtaining a second total job data volume of all jobs executed by multiple computing clusters. In this embodiment, aggregation can be performed according to different computing clusters, different user identifiers, and different resource types. The second total job data volume includes, but is not limited to, the sum of resource consumption, job volume, and storage volume for each job executed under the billing metrics. At least one of the billing metrics may include at least one of cluster identifier, user identifier, resource type, and resource unit price.

[0112] S860 determines the real-time job data in multiple computing clusters based on the total amount of the first job data and the total amount of the second job data.

[0113] Optionally, combined Figure 9 As shown, the cloud platform 102 can compare the total amount of the first job data with the total amount of the second job data to generate a comparison result of the total amount of job data; alternatively, the cloud platform 102 can also summarize the total amount of the first job data with the total amount of the second job data to generate a summary result. This comparison can be made by subtracting the total amount of the first job data from the total amount of the second job data; or by finding the quotient of the total amount of the first job data from the total amount of the second job data; or by determining whether the total amount of the first job data is equal to the total amount of the second job data. Of course, this embodiment does not limit the method of comparing the total amount of the first job data with the total amount of the second job data. The comparison result of the total amount of job data refers to the result obtained after comparing the total amount of the first job data with the total amount of the second job data. The summary result refers to the result obtained after summing the total amount of the first job data with the total amount of the second job data. Afterwards, the cloud platform 102 can determine the real-time job data in multiple computing clusters based on the comparison result or the summary result of the total amount of job data.

[0114] In this embodiment, within a preset time period, job data for each job executed by multiple computing clusters under the billing indicators can be obtained relatively accurately from multiple computing clusters according to the billing indicators. Based on the job data for each job, the total first job data volume for all jobs executed by multiple computing clusters within the preset time period can be calculated relatively accurately. Furthermore, a second job data volume for all jobs executed by multiple computing clusters within the preset time period can be obtained relatively accurately from multiple computing clusters according to at least one of the billing indicators. Then, based on the relatively accurate first job data volume and the relatively accurate second job data volume, the real-time job data in multiple computing clusters with high complexity and large data volume can be determined relatively accurately.

[0115] The above embodiments involve determining real-time job data in multiple computing clusters based on the total amount of the first job data and the total amount of the second job data. The specific method is described below. In one embodiment, such as... Figure 10 As shown, S860 includes:

[0116] S1020: Compare the total amount of data from the first task with the total amount of data from the second task, and generate a comparison result.

[0117] Optionally, such as Figure 11 As shown, Figure 11This is a flowchart illustrating the processing of abnormal job data in one embodiment. S1102, the cloud platform 102 can compare the total amount of first job data (all jobs executed by multiple computing clusters within a preset time period) with the total amount of second job data (all jobs executed by multiple computing clusters within the preset time period) to generate a comparison result. This comparison can be made by subtracting the total amount of first job data from the total amount of second job data; or by finding the quotient of the total amount of first job data from the total amount of second job data; or by determining whether the total amount of first job data is equal to the total amount of second job data. Of course, this embodiment does not limit the method of comparing the total amount of first job data with the total amount of second job data. The comparison result refers to the result obtained after comparing the total amount of first job data with the total amount of second job data. For example, assuming this embodiment compares by determining whether the total amount of first job data is equal to the total amount of second job data, then the comparison result is either that the total amount of first job data is equal to the total amount of second job data, or that the total amount of first job data is not equal to the total amount of second job data.

[0118] S1040, if the comparison result shows that the total amount of data for the first job is equal to the total amount of data for the second job, then the total amount of data for the first job or the total amount of data for the second job will be used as the real-time job data in multiple computing clusters.

[0119] Optionally, combined Figure 11 As shown, if the comparison result shows that the total amount of data for the first job is equal to the total amount of data for the second job, the comparison is considered successful. The cloud platform 102 can then confirm that both the total amount of data for the first job and the total amount of data for the second job are accurate. At this point, in step S1104, the cloud platform 102 records the user, time, comparison result, and real-time job data of the successful comparison. In step S1106, the cloud platform 102 can use either the total amount of data for the first job or the total amount of data for the second job as the real-time job data for multiple computing clusters. Generally, since the total amount of data for the second job is the detailed job data obtained in real-time from all jobs executed by multiple computing clusters, it is usually selected as the real-time job data for multiple computing clusters.

[0120] In this embodiment, the total amount of the first job data is compared with the total amount of the second job data to generate a comparison result. If the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data, it indicates that the total amount of job data transmitted from multiple computing clusters to the cloud platform is accurate. Therefore, either the total amount of the first job data or the total amount of the second job data can be used as the real-time job data in multiple computing clusters, thus determining relatively accurate real-time job data in multiple computing clusters with high complexity and large data volumes. Therefore, this embodiment can obtain job data for each group of users in each computing cluster for different resource types based on the dimension of multiple computing clusters, and periodically compare it with the detailed job data obtained from the cloud platform, thereby ensuring the consistency between the job data of multiple computing clusters and the cloud platform.

[0121] In the above embodiment, if the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data, then the total amount of the first job data or the total amount of the second job data is used as the real-time job data in multiple computing clusters. The following describes another specific method. In one embodiment, such as... Figure 12 As shown, the above reconciliation methods include:

[0122] S1220, if the comparison result shows that the total amount of data for the first job is not equal to the total amount of data for the second job, then within a preset time period, the job data for each job executed by multiple computing clusters is re-acquired, and based on the job data for each job, the new total amount of data for the first job executed by multiple computing clusters within the preset time period is calculated.

[0123] For example, in combination Figure 11 As shown, if the comparison result shows that the total amount of data for the first job is not equal to the total amount of data for the second job, it indicates that the comparison has failed. In this case, cloud platform 102 can determine that an error occurred during the transmission of the total amount of data for the first job and / or the total amount of data for the second job. At this time, in step S1108, cloud platform 102 can record the user, time, comparison result, and real-time job data of the failed comparison in the cloud platform database, and query the time when the comparison was last successful for that user. The time when the comparison was last successful refers to the time when the comparison was closest to the current time. Then, in step S1110, cloud platform 102 can determine the time difference between the user's last successful comparison time and the current time, thereby obtaining the time difference period corresponding to this time difference. In step S1112, cloud platform 102 can re-acquire the job data of each job executed by multiple computing clusters within the time difference period, and based on the job data of each job, calculate the new total amount of data for the first job executed by multiple computing clusters within the time difference period. The time difference period can be equal to a preset time period.

[0124] S1240, according to at least one of the billing metrics, reacquire the total amount of new second job data for executing all jobs on multiple computing clusters within a preset time period.

[0125] Optionally, combined Figure 11 As shown in step S1112, the cloud platform 102 can retrieve new detailed job data for all jobs executed by multiple computing clusters from multiple computing clusters via message queues within a preset time period. This detailed job data may include, but is not limited to, job execution duration, resource consumption, and queues used during job execution. Furthermore, the cloud platform 102 can aggregate the new detailed job data within the preset time period according to at least one of the billing metrics, thereby obtaining a new second set of job data for all jobs executed by multiple computing clusters.

[0126] S1260, based on the new total amount of the first job data and the new total amount of the second job data, determines the new real-time job data in multiple computing clusters.

[0127] Optionally, combined Figure 11 As shown, cloud platform 102 can compare the new total amount of first job data with the new total amount of second job data to generate a comparison result of the new total amount of job data. Then, cloud platform 102 can determine the new real-time job data in multiple computing clusters based on the comparison result of the new total amount of job data. Furthermore, in step S1114, cloud platform 102 can also detect whether billing has been initiated for the new total amount of first job data or the new total amount of second job data. If billing has already been initiated for the new total amount of first job data or the new total amount of second job data, then in step S1116, cloud platform 102 can record the billed total amount of job data.

[0128] If no billing is made for the new total amount of first or second job data, then in step S1118, the cloud platform 102 can bill for the new total amount of first or second job data and reconcile the billed new total amount of first or second job data to generate a reconciliation result. Alternatively, if the reconciliation result is a failure, the cloud platform 102 can re-acquire the new total amount of first and second job data and reconcile the data based on these new total amounts to generate a new reconciliation result. If the reconciliation results fail a preset number of times, the reconciliation process stops, and each job data is manually checked. Of course, this embodiment does not limit the preset number of times.

[0129] In this embodiment, if the comparison result shows that the total amount of the first job data is not equal to the total amount of the second job data, it indicates that an error occurred during the transmission of the total amount of the first job data and / or the total amount of the second job data during the transmission of the first job data and / or the second job data during the transmission of the first job data and / or the second job data during the transmission of the first job data and the second job data during the transmission of the second job data within the transmission of the second job data. Therefore, the new real-time job data in the multiple computing clusters can be re-determined based on the re-acquired total amount of the first job data and the re-acquired total amount of the second job data. Thus, this embodiment proposes a method for data processing in cases where job data is abnormal (i.e., inconsistencies in job data occur during reconciliation).

[0130] In an optional embodiment, such as Figure 13 As shown, a reconciliation method is provided, applied to cloud platform 102, including:

[0131] S1302, within a preset time period, acquire job data of each job executed by multiple computing clusters under the billing indicators, and calculate the total first job data of all jobs executed by multiple computing clusters within the preset time period based on the job data of each job.

[0132] S1304, according to at least one of the billing indicators, obtain the total amount of second job data for all jobs executed by multiple computing clusters within a preset time period; the total amount of second job data includes the sum of resource consumption for each job executed under the billing indicator;

[0133] S1306, compare the total amount of data from the first task with the total amount of data from the second task, and generate a comparison result;

[0134] S1308, if the comparison result is that the total amount of data for the first job is equal to the total amount of data for the second job, then the total amount of data for the first job or the total amount of data for the second job will be used as the real-time job data in multiple computing clusters.

[0135] S1310, if the comparison result is that the total amount of data for the first job is not equal to the total amount of data for the second job, then within a preset time period, the job data for each job executed by multiple computing clusters is re-acquired, and based on the job data for each job, the new total amount of data for the first job executed by multiple computing clusters within the preset time period is calculated.

[0136] S1312, according to at least one of the billing indicators, reacquire the new total amount of second job data for executing all jobs on multiple computing clusters within a preset time period;

[0137] S1314, Based on the new total amount of the first job data and the new total amount of the second job data, determine the new real-time job data in multiple computing clusters;

[0138] S1316, Aggregate real-time job data from multiple computing clusters according to the first type of aggregation index to generate the first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes billing index and the total resource consumption corresponding to the billing index; the billing index includes cluster identifier, user identifier, resource type and resource unit price;

[0139] S1318, Aggregate real-time job data in multiple computing clusters according to the second type of aggregation index to generate the second aggregation result corresponding to the second type of aggregation index; the second type of aggregation index includes billing index, resource consumption for each job executed under the billing index, and job cost data for each job executed under the billing index.

[0140] S1320, determine the total resource consumption corresponding to the same resource type and the same resource unit price from the first aggregation result, and calculate the total first operation cost corresponding to the same resource type and the same resource unit price based on the total resource consumption and the resource unit price;

[0141] S1322, Determine the operation cost data for each operation under the same resource type and the same resource unit price from the second aggregation result, and determine the total second operation cost based on the operation cost data for each operation;

[0142] S1324, If the total cost of the first operation is equal to the total cost of the second operation, then the reconciliation result is determined to be successful.

[0143] S1326, If the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, obtain the total first resource consumption corresponding to the billing indicator from the first aggregation result, and obtain the total second resource consumption corresponding to the billing indicator from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result.

[0144] S1328 compares the total first resource consumption and the total second resource consumption corresponding to each billing indicator to generate reconciliation results.

[0145] Optionally, such as Figure 14 As shown, Figure 14This is a schematic diagram of the overall process for reconciling real-time job data across multiple clusters in one embodiment. First, in step S1402, the user inputs the "Start Reconciliation" command on the front-end interface. In step S1404, the user inputs a preset time period on the front-end interface. Next, after receiving the "Start Reconciliation" command and the preset time period, the cloud platform obtains the total first and second job data volumes across multiple computing clusters within the preset time period through the job data verification module. The platform then compares the first and second job data volumes to generate a comparison result. In step S1406, if the comparison result shows that the first and second job data volumes are not equal, it indicates that the job data verification has failed. At this point, the abnormal job data processing module re-obtains the new first and second job data volumes from the multiple computing clusters within the preset time period and re-verifies the job data to generate new real-time job data.

[0146] S1408, If the comparison result shows that the total amount of data for the first job is equal to the total amount of data for the second job, it indicates that the job data verification is successful. At this time, the total amount of data for the first job or the total amount of data for the second job is used as the real-time job data in multiple computing clusters. S1410, Based on the real-time job data in multiple computing clusters, automatic reconciliation is performed through the cloud platform reconciliation module to generate a reconciliation result. S1412, If the reconciliation result is successful, the real-time job data and the message "Reconciliation Successful" are displayed on the front-end interface. S1414, If the reconciliation result is unsuccessful, the abnormal job data processing module re-obtains the new total amount of data for the first job and the new total amount of data for the second job in multiple computing clusters within the preset time period, and performs job data verification and reconciliation again to generate new real-time job data and a new reconciliation result.

[0147] The aforementioned reconciliation method aggregates real-time job data from multiple computing clusters according to different aggregation metrics. This allows for the classification and statistical analysis of highly complex, large-volume real-time job data from multiple computing clusters, generating accurate aggregation results for different metrics across multiple computing clusters. Subsequently, based on these accurate aggregation results for different metrics across multiple computing clusters, reconciliation of the highly complex, large-volume real-time job data from multiple computing clusters can be performed accurately, resulting in a more precise reconciliation outcome. Therefore, this application can reconcile real-time job data from multiple highly complex, large-volume computing clusters based on different aggregation metrics, achieving reconciliation across different dimensions and ensuring the accuracy of the job data. Furthermore, a front-end interface ensures users can view the job data and intuitively track fund flows. Additionally, an abnormal job data processing module records users and job data related to reconciliation failures; and when multiple reconciliation failures require manual intervention, the scope of the failures can be provided, reducing the difficulty of investigation.

[0148] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0149] Based on the same inventive concept, this application also provides a reconciliation device for implementing the reconciliation method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more reconciliation device embodiments provided below can be found in the limitations of the reconciliation method above, and will not be repeated here.

[0150] In one embodiment, such as Figure 15 As shown, a reconciliation device 1500 is provided, including: an acquisition module 1520, an aggregation module 1540, and a reconciliation module 1560, wherein:

[0151] The acquisition module 1520 is used to acquire real-time job data from multiple computing clusters.

[0152] The aggregation module 1540 is used to aggregate real-time job data from multiple computing clusters according to different aggregation metrics, and generate aggregation results corresponding to different aggregation metrics.

[0153] The reconciliation module 1560 is used to reconcile real-time job data in multiple computing clusters based on the aggregation results corresponding to different aggregation metrics, and generate reconciliation results.

[0154] In one embodiment, the different aggregation metrics include a first type of aggregation metric and a second type of aggregation metric; the aggregation module 1540 includes:

[0155] The first aggregation result generation unit is used to aggregate real-time job data from multiple computing clusters according to the first type of aggregation indicators to generate the first aggregation result corresponding to the first type of aggregation indicators; the first type of aggregation indicators includes billing indicators and the total resource consumption corresponding to the billing indicators; the billing indicators include cluster identifier, user identifier, resource type and resource unit price;

[0156] The second aggregation result generation unit is used to aggregate real-time job data in multiple computing clusters according to the second type of aggregation indicators to generate the second aggregation result corresponding to the second type of aggregation indicators. The second type of aggregation indicators includes billing indicators, resource consumption for each job executed under the billing indicators, and job cost data for each job executed under the billing indicators.

[0157] In one embodiment, the reconciliation module 1560 includes:

[0158] The first operation cost total calculation unit is used to determine the total resource consumption corresponding to the same resource type and the same resource unit price from the first aggregation result, and calculate the total first operation cost corresponding to the same resource type and the same resource unit price based on the total resource consumption and the resource unit price.

[0159] The second task cost determination unit is used to determine the task cost data for each task under the same resource type and the same resource unit price from the second aggregation result, and to determine the second task cost total based on the task cost data for each task.

[0160] The reconciliation result generation unit is used to reconcile real-time job data in multiple computing clusters based on the total cost of the first job and the total cost of the second job, and generate reconciliation results.

[0161] In one embodiment, generating reconciliation results includes:

[0162] The first reconciliation result determination subunit is used to determine the reconciliation result as successful when the total cost of the first operation is equal to the total cost of the second operation.

[0163] The resource consumption total acquisition subunit is used to, when the total first operation cost and the total second operation cost are not equal, obtain the first resource consumption total corresponding to each billing indicator from the first aggregation result and the second resource consumption total corresponding to each billing indicator from the second aggregation result; the second resource consumption total is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result.

[0164] The second reconciliation result determination subunit is used to compare the total first resource consumption and the total second resource consumption corresponding to each billing indicator and generate reconciliation results.

[0165] In one embodiment, the acquisition module 1520 includes:

[0166] The first task data total calculation unit is used to acquire task data of multiple computing clusters executing each task under the billing indicators within a preset time period, and calculate the first task data total of all tasks executed by multiple computing clusters within the preset time period based on the task data of each task.

[0167] The second job data total acquisition unit is used to acquire the second job data total of all jobs executed by multiple computing clusters within a preset time period according to at least one of the billing indicators; the second job data total includes the sum of resource consumption for each job executed under the billing indicators;

[0168] The real-time job data acquisition unit is used to determine the real-time job data in multiple computing clusters based on the total amount of the first job data and the total amount of the second job data.

[0169] In one embodiment, the real-time job data acquisition unit includes:

[0170] The comparison unit is used to compare the total amount of data in the first job with the total amount of data in the second job and generate a comparison result.

[0171] The real-time job data determination unit is used to determine the total amount of the first job data or the total amount of the second job data as the real-time job data in multiple computing clusters when the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data.

[0172] In one embodiment, the reconciliation device 1500 further includes:

[0173] The module for re-acquiring the total amount of first job data is used to re-acquire the job data of each job executed by multiple computing clusters within a preset time period when the comparison result shows that the total amount of first job data is not equal to the total amount of second job data. Based on the job data of each job, the module calculates the new total amount of first job data of all jobs executed by multiple computing clusters within the preset time period.

[0174] The second job data total reacquisition module is used to reacquire the new second job data total for all jobs executed by multiple computing clusters within a preset time period according to at least one of the billing indicators.

[0175] A new real-time job data generation module is used to determine new real-time job data in multiple computing clusters based on the new total amount of first job data and the new total amount of second job data.

[0176] Each module in the aforementioned reconciliation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0177] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 16 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores reconciliation data. The I / O interfaces allow the processor to exchange information with external devices. The communication interface allows communication with external terminals via a network connection. When executed by the processor, the computer program implements a reconciliation method.

[0178] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0179] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0180] Acquire real-time job data from multiple computing clusters;

[0181] Real-time job data from multiple computing clusters are aggregated according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics.

[0182] Based on the aggregation results corresponding to different aggregation metrics, the real-time job data in multiple computing clusters are reconciled to generate reconciliation results.

[0183] In one embodiment, the different aggregation metrics include a first type of aggregation metric and a second type of aggregation metric; real-time job data from multiple computing clusters are aggregated according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics; when the processor executes the computer program, it also implements the following steps:

[0184] The real-time job data from multiple computing clusters are aggregated according to the first type of aggregation index to generate the first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes the billing index and the total resource consumption corresponding to the billing index; the billing index includes the cluster identifier, user identifier, resource type and resource unit price;

[0185] The real-time job data from multiple computing clusters are aggregated according to the second type of aggregation index to generate the second aggregation result corresponding to the second type of aggregation index. The second type of aggregation index includes billing index, resource consumption for each job executed under the billing index, and job cost data for each job executed under the billing index.

[0186] In one embodiment, real-time job data in multiple computing clusters are reconciled based on the aggregation results corresponding to different aggregation metrics to generate reconciliation results. When the processor executes the computer program, it also performs the following steps:

[0187] Determine the total resource consumption corresponding to the same resource type and the same resource unit price from the first aggregation result. Calculate the total first operation cost corresponding to the same resource type and the same resource unit price based on the total resource consumption and the resource unit price.

[0188] From the second aggregation results, determine the operation cost data for each operation under the same resource type and the same resource unit price, and determine the total second operation cost based on the operation cost data for each operation;

[0189] Based on the total cost of the first job and the total cost of the second job, the real-time job data in multiple computing clusters are reconciled to generate reconciliation results.

[0190] In one embodiment, real-time job data across multiple computing clusters are reconciled based on the total first job cost and the total second job cost to generate a reconciliation result. The processor, when executing the computer program, also performs the following steps:

[0191] If the total cost of the first task is equal to the total cost of the second task, then the reconciliation result is determined to be successful.

[0192] If the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result.

[0193] Compare the total first resource consumption and the total second resource consumption corresponding to each billing indicator to generate reconciliation results.

[0194] In one embodiment, real-time job data from multiple computing clusters is acquired, and the processor, while executing a computer program, further performs the following steps:

[0195] Within a preset time period, acquire job data for each job executed by multiple computing clusters under the billing indicators, and calculate the total first job data of all jobs executed by multiple computing clusters within the preset time period based on the job data of each job.

[0196] Based on at least one of the billing metrics, the total amount of second job data is obtained within a preset time period for executing all jobs across multiple computing clusters; the total amount of second job data includes the sum of resource consumption for each job executed under the billing metrics.

[0197] Based on the total amount of data from the first job and the total amount of data from the second job, the real-time job data in multiple computing clusters is determined.

[0198] In one embodiment, real-time job data in multiple computing clusters is determined based on the total amount of first job data and the total amount of second job data. When the processor executes the computer program, it further performs the following steps:

[0199] Compare the total amount of data from the first task with the total amount of data from the second task, and generate a comparison result;

[0200] If the comparison result shows that the total amount of data for the first job is equal to the total amount of data for the second job, then the total amount of data for the first job or the total amount of data for the second job will be used as the real-time job data in multiple computing clusters.

[0201] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0202] If the comparison result shows that the total amount of data for the first job is not equal to the total amount of data for the second job, then within a preset time period, the job data for each job executed by multiple computing clusters will be reacquired, and based on the job data for each job, the new total amount of data for the first job executed by multiple computing clusters within the preset time period will be calculated.

[0203] Based on at least one of the billing metrics, reacquire the total amount of new second job data for executing all jobs across multiple computing clusters within a preset time period;

[0204] Based on the total amount of new first job data and the total amount of new second job data, new real-time job data is determined in multiple computing clusters.

[0205] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0206] Acquire real-time job data from multiple computing clusters;

[0207] Real-time job data from multiple computing clusters are aggregated according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics.

[0208] Based on the aggregation results corresponding to different aggregation metrics, the real-time job data in multiple computing clusters are reconciled to generate reconciliation results.

[0209] In one embodiment, the different aggregation metrics include a first type of aggregation metrics and a second type of aggregation metrics; real-time job data from multiple computing clusters are aggregated according to different aggregation metrics to generate aggregation results corresponding to different aggregation metrics; when the computer program is executed by the processor, the following steps are also implemented:

[0210] The real-time job data from multiple computing clusters are aggregated according to the first type of aggregation index to generate the first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes the billing index and the total resource consumption corresponding to the billing index; the billing index includes the cluster identifier, user identifier, resource type and resource unit price;

[0211] The real-time job data from multiple computing clusters are aggregated according to the second type of aggregation index to generate the second aggregation result corresponding to the second type of aggregation index. The second type of aggregation index includes billing index, resource consumption for each job executed under the billing index, and job cost data for each job executed under the billing index.

[0212] In one embodiment, real-time job data in multiple computing clusters are reconciled based on the aggregation results corresponding to different aggregation metrics to generate reconciliation results. When the computer program is executed by the processor, it also performs the following steps:

[0213] Determine the total resource consumption corresponding to the same resource type and the same resource unit price from the first aggregation result. Calculate the total first operation cost corresponding to the same resource type and the same resource unit price based on the total resource consumption and the resource unit price.

[0214] From the second aggregation results, determine the operation cost data for each operation under the same resource type and the same resource unit price, and determine the total second operation cost based on the operation cost data for each operation;

[0215] Based on the total cost of the first job and the total cost of the second job, the real-time job data in multiple computing clusters are reconciled to generate reconciliation results.

[0216] In one embodiment, real-time job data across multiple computing clusters are reconciled based on the total first job cost and the total second job cost to generate a reconciliation result. When the computer program is executed by the processor, it also performs the following steps:

[0217] If the total cost of the first task is equal to the total cost of the second task, then the reconciliation result is determined to be successful.

[0218] If the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result.

[0219] Compare the total first resource consumption and the total second resource consumption corresponding to each billing indicator to generate reconciliation results.

[0220] In one embodiment, real-time job data from multiple computing clusters is acquired, and the computer program, when executed by a processor, further performs the following steps:

[0221] Within a preset time period, acquire job data for each job executed by multiple computing clusters under the billing indicators, and calculate the total first job data of all jobs executed by multiple computing clusters within the preset time period based on the job data of each job.

[0222] Based on at least one of the billing metrics, the total amount of second job data is obtained within a preset time period for executing all jobs across multiple computing clusters; the total amount of second job data includes the sum of resource consumption for each job executed under the billing metrics.

[0223] Based on the total amount of data from the first job and the total amount of data from the second job, the real-time job data in multiple computing clusters is determined.

[0224] In one embodiment, real-time job data in multiple computing clusters is determined based on the total amount of first job data and the total amount of second job data. When the computer program is executed by the processor, the following steps are also performed:

[0225] Compare the total amount of data from the first task with the total amount of data from the second task, and generate a comparison result;

[0226] If the comparison result shows that the total amount of data for the first job is equal to the total amount of data for the second job, then the total amount of data for the first job or the total amount of data for the second job will be used as the real-time job data in multiple computing clusters.

[0227] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0228] If the comparison result shows that the total amount of data for the first job is not equal to the total amount of data for the second job, then within a preset time period, the job data for each job executed by multiple computing clusters will be reacquired, and based on the job data for each job, the new total amount of data for the first job executed by multiple computing clusters within the preset time period will be calculated.

[0229] Based on at least one of the billing metrics, reacquire the total amount of new second job data for executing all jobs across multiple computing clusters within a preset time period;

[0230] Based on the total amount of new first job data and the total amount of new second job data, new real-time job data is determined in multiple computing clusters.

[0231] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0232] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0233] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0234] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0235] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A reconciliation method, characterized in that, Applied to a cloud platform, the method includes: Acquire real-time job data from multiple computing clusters; The real-time job data in the multiple computing clusters are classified and merged according to different aggregation metrics to generate aggregation results corresponding to the different aggregation metrics; the aggregation results include a job aggregation table and an account transaction table. Based on the aggregation results corresponding to the different aggregation indicators, determine the product of the total resource consumption and the corresponding resource unit price for the same resource type and the same resource unit price in the real-time job data, as well as the sum of the job cost data for each job for the same resource type and the same resource unit price in the real-time job data, and reconcile the real-time job data in the multiple computing clusters based on the product and the sum to generate a reconciliation result; The step of reconciling the real-time job data in the multiple computing clusters based on the product and the sum to generate a reconciliation result includes: if the product and the sum are not equal, comparing the first total resource consumption and the second total resource consumption corresponding to each billing indicator to generate the reconciliation result.

2. The method according to claim 1, characterized in that, The different aggregation metrics include a first type of aggregation metrics and a second type of aggregation metrics; the process of classifying and merging real-time job data in the multiple computing clusters according to different aggregation metrics to generate aggregation results corresponding to the different aggregation metrics includes: The real-time job data in the multiple computing clusters are aggregated according to the first type of aggregation index to generate a first aggregation result corresponding to the first type of aggregation index; the first type of aggregation index includes a billing index and the total resource consumption corresponding to the billing index; the billing index includes a cluster identifier, a user identifier, a resource type, and a resource unit price. The real-time job data in the multiple computing clusters are aggregated according to the second type of aggregation index to generate a second aggregation result corresponding to the second type of aggregation index; the second type of aggregation index includes the billing index, the resource consumption of each job executed under the billing index, and the job cost data of each job executed under the billing index.

3. The method according to claim 2, characterized in that, The process involves determining the product of the total resource consumption and the corresponding resource unit price in the real-time job data, and the sum of the job cost data for each job in the real-time job data, based on the aggregation results corresponding to the different aggregation indicators. Then, the real-time job data in the multiple computing clusters is reconciled based on the product and the sum to generate a reconciliation result, including: The total resource consumption corresponding to the same resource type and the same resource unit price is determined from the first aggregation result. Based on the total resource consumption and the resource unit price, the total first operation cost corresponding to the same resource type and the same resource unit price is calculated. From the second aggregation result, determine the operation cost data for each operation under the same resource type and the same resource unit price, and determine the total second operation cost based on the operation cost data for each operation; Based on the total cost of the first job and the total cost of the second job, the real-time job data in the multiple computing clusters are reconciled to generate the reconciliation result.

4. The method according to claim 3, characterized in that, The step of reconciling the real-time job data in the multiple computing clusters based on the total first job cost and the total second job cost, and generating the reconciliation result, includes: If the total cost of the first operation is equal to the total cost of the second operation, then the reconciliation result is determined to be a successful reconciliation. If the total cost of the first operation is not equal to the total cost of the second operation, then for each billing indicator, the total first resource consumption corresponding to the billing indicator is obtained from the first aggregation result, and the total second resource consumption corresponding to the billing indicator is obtained from the second aggregation result; the total second resource consumption is the sum of the resource consumption of each operation executed under the billing indicator in the second aggregation result; The total first resource consumption and the total second resource consumption corresponding to each of the aforementioned billing indicators are compared to generate the reconciliation result.

5. The method according to any one of claims 2-4, characterized in that, The acquisition of real-time job data from multiple computing clusters includes: Within a preset time period, the job data of each job executed by the multiple computing clusters under the billing indicators is obtained, and based on the job data of each job, the first job data total of all jobs executed by the multiple computing clusters within the preset time period is calculated. According to at least one of the billing metrics, the total amount of second job data for all jobs executed by the plurality of computing clusters within the preset time period is obtained; the total amount of second job data includes the sum of the resource consumption for each job executed under the billing metrics; Based on the total amount of the first job data and the total amount of the second job data, the real-time job data in the plurality of computing clusters is determined.

6. The method according to claim 5, characterized in that, The step of determining the real-time job data in the plurality of computing clusters based on the total amount of the first job data and the total amount of the second job data includes: Compare the total amount of the first task data with the total amount of the second task data to generate a comparison result; If the comparison result shows that the total amount of the first job data is equal to the total amount of the second job data, then the total amount of the first job data or the total amount of the second job data shall be used as the real-time job data in the plurality of computing clusters.

7. The method according to claim 6, characterized in that, The method further includes: If the comparison result is that the total amount of the first job data is not equal to the total amount of the second job data, then within the preset time period, the job data of each job executed by the multiple computing clusters is re-acquired, and based on the job data of each job, the new total amount of the first job data of all jobs executed by the multiple computing clusters within the preset time period is calculated. According to at least one of the billing metrics, within the preset time period, reacquire the total amount of new second job data for all jobs executed by the multiple computing clusters. Based on the total amount of the new first job data and the total amount of the new second job data, new real-time job data is determined in the plurality of computing clusters.

8. A reconciliation device, characterized in that, The device, applied to a cloud platform, includes: The acquisition module is used to acquire real-time job data from multiple computing clusters; The aggregation module is used to classify and merge real-time job data in the multiple computing clusters according to different aggregation metrics, and generate aggregation results corresponding to the different aggregation metrics; the aggregation results include a job aggregation table and an account transaction table. The reconciliation module is used to determine, based on the aggregation results corresponding to the different aggregation indicators, the product of the total resource consumption and the corresponding resource unit price for the resource type and the real-time job data, as well as the sum of the job cost data for each job corresponding to the same resource type and the same resource unit price in the real-time job data, and to reconcile the real-time job data in the multiple computing clusters based on the product and the sum to generate a reconciliation result; the reconciliation module includes: if the product and the sum are not equal, comparing the first total resource consumption and the second total resource consumption corresponding to each billing indicator to generate the reconciliation result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images