Fault supervision method and system for coping with complex scene based on flink framework

Abstract

The invention discloses a fault supervision method and system for coping with a complex scene based on an flink framework, and relates to the technical field of fault monitoring, and the method comprises the steps: receiving a real-time alarm message, carrying out rule matching, priority and mutual exclusion processing, data assembly and initialization, and putting the information into a supervision waiting queue; performing regular polling on the supervision waiting queue, automatically selecting supervision personnel for supervision tasks meeting rules, automatically performing replacement and desensitization processing on fault key information in the supervision tasks according to preset general configuration, and merging and compressing faults with the same characteristics at the same time; and finally, integrating the processed supervision tasks, and sending the integrated supervision tasks to a specified topic. Reading topic, calling a corresponding supervision interface according to the supervision mode of the supervision task, and executing actual supervision operation; and receiving a supervision result of executing the supervision task, and updating the supervision waiting queue. The invention aims to conveniently, quickly and accurately supervise faults in real time and quickly help maintenance personnel to discover and solve fault problems in time.

Description

Technical Field

[0001] The present invention relates to the technical field of fault monitoring, and in particular to a fault supervision method and system for coping with complex scenarios based on a flink framework. Background Art

[0002] Operator fault supervision relies on a large number of supervision rules and special processing to issue alarms and dispatch orders. It has poor adaptability to business changes, wastes a lot of human resources, and maintenance costs increase year by year.

[0003] With the development of the big data era, the number of alarms is increasing day by day, and business needs are becoming more and more complex. The traditional supervision operation model cannot meet the needs of development. First, the number of alarm fields increases year by year, and more and more field information is given business attributes. The business information between different professions is also different. The traditional method of using fixed fields to match conditions will inevitably be eliminated. Second, relying solely on supervision rules can no longer completely cover all the supervised businesses. It is necessary to cooperate with special processing to meet business needs, but it requires a lot of R&D costs and is not easy to maintain. Third, in the face of the surge in data volume, the traditional program method is limited by the processing performance of the physical machine and is facing increasing processing pressure. Fourth, it is not easy to find the relationship between each alarm in the traditional method. The same fault will lead to multiple supervisions, repeated supervisions, and missed supervisions, which increases the difficulty of fault handling. Summary of the invention

[0004] In view of the needs and shortcomings of current technological development, the present invention provides a fault supervision method and system for dealing with complex scenarios based on the Flink framework, aiming to more conveniently, quickly and accurately supervise faults in real time, and quickly help maintenance personnel to promptly discover and solve fault problems.

[0005] In the first aspect, the present invention provides a fault supervision method for dealing with complex scenarios based on the flink framework. The technical solution adopted to solve the above technical problems is as follows:

[0006] A fault supervision method for complex scenarios based on the Flink framework includes the following steps:

[0007] S1. Use Kafka as a message queue to receive real-time alarm messages;

[0008] S2, perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm message, and put it into the supervision waiting queue;

[0009] S3. Regularly poll the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, automatically select the specific supervisor, and automatically replace and desensitize the key fault information in the supervision task according to the preset general configuration. At the same time, merge and compress faults with the same characteristics, and finally integrate the processed supervision tasks and send them to the designated supervision topic;

[0010] S4. Read the supervision task in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation;

[0011] S5. Receive the supervision result of the completed supervision task to update the status of the corresponding supervision task in the supervision waiting queue.

[0012] Optionally, step S2 is executed to perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm message, and put it into the supervision waiting queue, which specifically includes the following operations:

[0013] Use JEXL expressions to match alarm messages with preset suppression supervision rules, remove matched alarm messages from the message queue, and do not process them;

[0014] Use JEXL expressions to match alarm messages with preset supervision rules. For alarm messages that match supervision rules, determine the priority of the matched supervision rules based on the alarm business fields of the supervision rules. When there is a mutually exclusive relationship between the matched supervision rules, select the supervision rule with the highest priority.

[0015] According to the determined supervision rules, data is assembled in combination with specific alarm information, a supervision task is generated, placed in the supervision waiting queue, and initialized to a state of waiting for supervision.

[0016] Further optionally, the preset supervision rules include supervision type, alarm business field, field priority, supervision group, supervision status, supervision repetition times, supervision repetition time interval, supervision delay time and supervision area range;

[0017] When using JEXL expressions to match alarm messages with preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of these two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, automatically match the supervision group based on the supervision type, alarm business field, and selected field priority of the alarm message, and configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range.

[0018] Based on the above matching information and configuration information, data is assembled in combination with specific alarm information, a supervision task is generated, placed in the supervision waiting queue, and initialized to a state of waiting for supervision.

[0019] Further optionally, the supervision types involved include alarm, dispatch and work order, and the supervision waiting queue includes alarm supervision waiting queue, dispatch supervision waiting queue and work order supervision waiting queue; based on the supervision type of the alarm message, the corresponding generated supervision task is placed in the alarm supervision waiting queue, dispatch supervision waiting queue or work order supervision waiting queue;

[0020] The supervision status includes eight types: not supervised, waiting for supervision, in supervision, successful supervision, failed supervision, repeated supervision, resumed supervision and suppressed supervision. Among them, when the supervision waiting queue is initialized, the supervision tasks in the not supervised state are initialized to the waiting for supervision state.

[0021] Further optional step S3 is performed to periodically poll the supervision tasks in the supervision waiting queue:

[0022] For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area range, and then the subsequent operations will be performed;

[0023] For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the configured supervision group according to the supervision area range, and then subsequent operations are performed.

[0024] In the second aspect, the present invention provides a fault supervision system for dealing with complex scenarios based on the flink framework. The technical solutions adopted to solve the above technical problems are as follows:

[0025] A fault supervision system based on the Flink framework to deal with complex scenarios, including:

[0026] The flink real-time processing module is used to read the real-time alarm messages in the Kafka message queue, perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm messages, and put them into the supervision waiting queue;

[0027] The scheduled polling processing module is used to perform scheduled polling on the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, it automatically selects a specific supervisor, and automatically replaces and desensitizes the key fault information in the supervision task according to the preset general configuration. At the same time, faults with the same characteristics are merged and compressed, and finally the processed supervision tasks are integrated and sent to the specified supervision topic;

[0028] The real-time supervision module is used to read the supervision tasks in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation;

[0029] The supervision result acquisition module is used to receive the supervision results of the completed supervision tasks to update the status of the corresponding supervision tasks in the supervision waiting queue.

[0030] Optionally, the flink real-time processing modules involved include:

[0031] Message reading unit, used to read real-time alarm messages in the Kafka message queue;

[0032] A rule configuration unit, used to preset suppression supervision rules and supervision rules related to the alarm message;

[0033] Rule matching unit 1 is used to match the alarm message with the preset suppression supervision rule using JEXL expression, and remove the matched alarm message from the message queue without processing;

[0034] Rule matching unit 2 is used to match the alarm message with the preset supervision rule using JEXL expression, and for the alarm message matching the supervision rule, determine the priority of the matched supervision rule according to the alarm business field of the supervision rule, and select the supervision rule with the highest priority when there is a mutually exclusive relationship between the matched supervision rules;

[0035] The data assembly unit is used to assemble data according to the determined supervision rules and specific alarm information, generate supervision tasks, and put them into the supervision waiting queue;

[0036] The state initialization unit is used to initialize the supervision tasks in the supervision waiting queue to the state of waiting for supervision.

[0037] Further optionally, the supervision rules preset in the rule configuration unit involved include supervision type, alarm business field, field priority, supervision group, supervision status, supervision repetition times, supervision repetition time interval, supervision delay time and supervision area range;

[0038] Rule matching unit 2 uses JEXL expressions to match alarm messages with preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of these two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, automatically match the supervision group based on the supervision type, alarm business field, and selected field priority of the alarm message, and configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range.

[0039] The data assembly unit assembles data based on the above matching information and configuration information in combination with specific alarm information, generates supervision tasks, and puts them into the supervision waiting queue.

[0040] Further optionally, the supervision types involved include alarm, dispatch and work order, and the supervision waiting queue includes alarm supervision waiting queue, dispatch supervision waiting queue and work order supervision waiting queue; the data assembly unit puts the corresponding generated supervision task into the alarm supervision waiting queue, dispatch supervision waiting queue or work order supervision waiting queue based on the supervision type of the alarm message;

[0041] The supervision status includes eight types: unsupervised, waiting for supervision, in supervision, successful supervision, failed supervision, repeated supervision, resumed supervision and suppressed supervision. Among them, when the status initialization unit initializes the supervision waiting queue, the supervision task in the unsupervised state is initialized to the waiting for supervision state.

[0042] Further optionally, when the timing polling processing module performs timing polling on the supervision tasks in the supervision waiting queue:

[0043] For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area range, and then the subsequent operations will be performed;

[0044] For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the configured supervision group according to the supervision area range, and then subsequent operations are performed.

[0045] The present invention provides a fault supervision method and system for dealing with complex scenarios based on the flink framework, which has the following beneficial effects compared with the prior art:

[0046] The present invention aims to more conveniently, quickly and accurately supervise faults in real time, quickly help maintenance personnel to promptly discover and solve fault problems, and greatly reduce labor costs; by introducing the real-time processing engine flink, it can truly achieve high throughput and low latency, greatly reducing the time for processing data, and stably facing alarm storms. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Attached Figure 1 is a flow chart of a method according to Embodiment 1 of the present invention;

[0048] Attached Figure 2 is a specific flow chart of step S2 in the first embodiment of the present invention;

[0049] Attached Figure 3 This is a module connection block diagram of the second embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the technical solution, the technical problem solved and the technical effect of the present invention more clearly understood, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0051] Embodiment 1:

[0052] Combined with Figure 1 , 2 This embodiment proposes a fault supervision method for dealing with complex scenarios based on the Flink framework, which includes the following steps:

[0053] S1. Use Kafka as a message queue to receive real-time alarm messages.

[0054] S2: Perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm message, and put it into the supervision waiting queue, which specifically includes the following operations:

[0055] S2.1. Use JEXL expressions to match alarm messages with preset suppression supervision rules, remove the matched alarm messages from the message queue, and do not process them.

[0056] S2.2. The preset supervision rules include supervision type, alarm business field, field priority, supervision group, supervision status, supervision repetition times, supervision repetition time interval, supervision delay time and supervision area range. Among them, the supervision types include alarm, dispatch and work order. The supervision status includes unsupervised, waiting for supervision, in progress, successful supervision, failed supervision, repeated supervision, resumed supervision and suppressed supervision.

[0057] Use JEXL expressions to match alarm messages with preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of these two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, automatically match the supervision group based on the supervision type, alarm business field, and selected field priority of the alarm message, and configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range (such as cities, districts, counties, and maintenance companies).

[0058] S2.3. Based on the above matching information and configuration information, data is assembled in combination with specific alarm information to generate supervision tasks and put them into the supervision waiting queue, where the supervision waiting queue includes three types: alarm supervision waiting queue, dispatch supervision waiting queue or work order supervision waiting queue. Based on the supervision type, the supervision tasks are placed into the alarm supervision waiting queue, dispatch supervision waiting queue or work order supervision waiting queue, and the supervision tasks in the non-supervised state are initialized to the state waiting for supervision.

[0059] S3. Regularly poll the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the supervision group according to the supervision area (such as cities, districts, counties, and maintenance companies). The key fault information in the supervision task is automatically replaced and desensitized according to the preset general configuration. At the same time, faults with the same characteristics are merged and compressed. Finally, the processed supervision tasks are integrated and sent to the designated supervision topic.

[0060] It should be added that: when periodically polling the supervision tasks in the supervision waiting queue:

[0061] For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area range, and then the subsequent operations will be performed;

[0062] For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the configured supervision group according to the supervision area range, and then subsequent operations are performed.

[0063] S4. Read the supervision tasks in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation.

[0064] S5. Receive the supervision result of the completed supervision task to update the status of the corresponding supervision task in the supervision waiting queue.

[0065] Embodiment 2:

[0066] Combined with Figure 3 This embodiment proposes a fault supervision system for dealing with complex scenarios based on the flink framework, which includes:

[0067] The flink real-time processing module is used to read the real-time alarm messages in the Kafka message queue, perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm messages, and put them into the supervision waiting queue;

[0068] The scheduled polling processing module is used to perform scheduled polling on the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, it automatically selects a specific supervisor, and automatically replaces and desensitizes the key fault information in the supervision task according to the preset general configuration. At the same time, faults with the same characteristics are merged and compressed, and finally the processed supervision tasks are integrated and sent to the specified supervision topic;

[0069] The real-time supervision module is used to read the supervision tasks in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation;

[0070] The supervision result acquisition module is used to receive the supervision results of the completed supervision tasks to update the status of the corresponding supervision tasks in the supervision waiting queue.

[0071] In this embodiment, the flink real-time processing module involved also includes:

[0072] Message reading unit, used to read real-time alarm messages in the Kafka message queue;

[0073] A rule configuration unit, used to preset suppression supervision rules and supervision rules related to the alarm message;

[0074] Rule matching unit 1 is used to match the alarm message with the preset suppression supervision rule using JEXL expression, and remove the matched alarm message from the message queue without processing;

[0075] Rule matching unit 2 is used to match the alarm message with the preset supervision rule using JEXL expression, and for the alarm message matching the supervision rule, determine the priority of the matched supervision rule according to the alarm business field of the supervision rule, and select the supervision rule with the highest priority when there is a mutually exclusive relationship between the matched supervision rules;

[0076] The data assembly unit is used to assemble data according to the determined supervision rules and specific alarm information, generate supervision tasks, and put them into the supervision waiting queue;

[0077] The state initialization unit is used to initialize the supervision tasks in the supervision waiting queue to the state of waiting for supervision.

[0078] It should be specifically noted that the supervision rules preset in the rule configuration unit include supervision type, alarm business field, field priority, supervision group, supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time and supervision area range. Among them, the supervision types include alarm, dispatch and work order. The supervision waiting queue includes the alarm supervision waiting queue, dispatch supervision waiting queue and work order supervision waiting queue. The supervision status includes eight types: not supervised, waiting for supervision, in progress, supervision successful, supervision failed, repeated supervision, resumed supervision and suppressed supervision. The rule matching unit 2 involved uses JEXL expressions to match the alarm message with the preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of these two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, based on the supervision type, alarm business field, and selected field priority of the alarm message, automatically match the supervision group, configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range; the data assembly unit involved assembles data based on the above matching information and configuration information, combined with specific alarm information, generates supervision tasks, and puts them into the corresponding alarm supervision waiting queue, dispatch supervision waiting queue, or work order supervision waiting queue; when the state initialization unit initializes the supervision waiting queue, the supervision tasks in the non-supervised state are initialized to the state of waiting for supervision.

[0079] In this embodiment, when the timing polling processing module performs timing polling on the supervision tasks in the supervision waiting queue:

[0080] For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area (such as cities, districts, counties, and maintenance companies), and then the subsequent operations will be performed;

[0081] For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervision personnel are automatically selected from the configured supervision group according to the supervision area scope (such as cities, counties, and maintenance companies), and then subsequent operations are performed.

[0082] In summary, the fault supervision method and system based on the flink framework for dealing with complex scenarios of the present invention can conveniently, quickly and accurately supervise faults in real time, quickly help maintenance personnel to promptly discover and solve fault problems, and reduce labor costs.

[0083] The above specific examples are used to explain the principles and implementation methods of the present invention in detail. These examples are only used to help understand the core technical content of the present invention. Based on the above specific embodiments of the present invention, any improvements and modifications made by technicians in this technical field without departing from the principles of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A fault supervision method for complex scenarios based on the Flink framework, characterized in that: The steps include: S1. Use Kafka as a message queue to receive real-time alarm messages; S2, perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm message, and put it into the supervision waiting queue; S3. Regularly poll the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, automatically select the specific supervisor, and automatically replace and desensitize the key fault information in the supervision task according to the preset general configuration. At the same time, merge and compress faults with the same characteristics, and finally integrate the processed supervision tasks and send them to the designated supervision topic; S4. Read the supervision task in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation; S5. Receive the supervision result of the completed supervision task to update the status of the corresponding supervision task in the supervision waiting queue.

2. According to claim 1, a fault supervision method for dealing with complex scenarios based on the flink framework is characterized in that: Execute step S2 to perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm message, and put it into the supervision waiting queue, which specifically includes the following operations: Use JEXL expressions to match alarm messages with preset suppression supervision rules, remove matched alarm messages from the message queue, and do not process them; Use JEXL expressions to match alarm messages with preset supervision rules. For alarm messages that match supervision rules, determine the priority of the matched supervision rules based on the alarm business fields of the supervision rules. When there is a mutually exclusive relationship between the matched supervision rules, select the supervision rule with the highest priority. According to the determined supervision rules, data is assembled in combination with specific alarm information, a supervision task is generated, placed in the supervision waiting queue, and initialized to a state of waiting for supervision.

3. According to claim 2, a fault supervision method for complex scenarios based on the flink framework is characterized in that: The preset supervision rules include supervision type, alarm business field, field priority, supervision group, supervision status, supervision repetition times, supervision repetition time interval, supervision delay time and supervision area range; When using JEXL expressions to match alarm messages with preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of these two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, automatically match the supervision group based on the supervision type, alarm business field, and selected field priority of the alarm message, and configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range. Based on the above matching information and configuration information, data is assembled in combination with specific alarm information, a supervision task is generated, placed in the supervision waiting queue, and initialized to a state of waiting for supervision.

4. According to claim 3, a fault supervision method for dealing with complex scenarios based on the flink framework is characterized in that: The supervision types include alarm, dispatch and work order. The supervision waiting queue includes alarm supervision waiting queue, dispatch supervision waiting queue and work order supervision waiting queue. Based on the supervision type of the alarm message, the corresponding generated supervision task is placed in the alarm supervision waiting queue, dispatch supervision waiting queue or work order supervision waiting queue. The supervision status includes eight types: unsupervised, waiting for supervision, in supervision, successful supervision, failed supervision, repeated supervision, resumed supervision and suppressed supervision. When the supervision waiting queue is initialized, the supervision tasks in the unsupervised state are initialized to the waiting for supervision state.

5. According to claim 4, a fault supervision method for dealing with complex scenarios based on the flink framework is characterized in that: When executing step S3 and performing periodic polling on the supervision tasks in the supervision waiting queue: For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area range, and then the subsequent operations will be performed; For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the configured supervision group according to the supervision area range, and then subsequent operations are performed.

6. A fault supervision system based on the flink framework for dealing with complex scenarios, characterized by: It includes: The flink real-time processing module is used to read the real-time alarm messages in the Kafka message queue, perform rule matching, priority and mutual exclusion processing, data assembly and initialization on the alarm messages, and put them into the supervision waiting queue; The scheduled polling processing module is used to perform scheduled polling on the supervision tasks in the supervision waiting queue. When the supervision task meets the supervision delay time, it automatically selects a specific supervisor, and automatically replaces and desensitizes the key fault information in the supervision task according to the preset general configuration. At the same time, faults with the same characteristics are merged and compressed, and finally the processed supervision tasks are integrated and sent to the specified supervision topic; The real-time supervision module is used to read the supervision tasks in the supervision topic, call the corresponding supervision interface according to the supervision method of the supervision task, and perform the actual supervision operation; The supervision result acquisition module is used to receive the supervision results of the completed supervision tasks to update the status of the corresponding supervision tasks in the supervision waiting queue.

7. According to claim 6, a fault supervision system for dealing with complex scenarios based on the flink framework is characterized in that: The flink real-time processing module further includes: Message reading unit, used to read real-time alarm messages in the Kafka message queue; A rule configuration unit, used to preset suppression supervision rules and supervision rules related to the alarm message; Rule matching unit 1 is used to match the alarm message with the preset suppression supervision rule using JEXL expression, and remove the matched alarm message from the message queue without processing; Rule matching unit 2 is used to match the alarm message with the preset supervision rule using JEXL expression, and for the alarm message matching the supervision rule, determine the priority of the matched supervision rule according to the alarm business field of the supervision rule, and select the supervision rule with the highest priority when there is a mutually exclusive relationship between the matched supervision rules; The data assembly unit is used to assemble data according to the determined supervision rules and specific alarm information, generate supervision tasks, and put them into the supervision waiting queue; The state initialization unit is used to initialize the supervision tasks in the supervision waiting queue to the state of waiting for supervision.

8. According to claim 7, a fault supervision system for dealing with complex scenarios based on the flink framework is characterized in that: The supervision rules preset by the rule configuration unit include supervision type, alarm service field, field priority, supervision group, supervision status, supervision repetition times, supervision repetition time interval, supervision delay time and supervision area range; The rule matching unit 2 uses JEXL expressions to match the alarm message with the preset supervision rules: first match the supervision type, then further match the alarm business field based on the matched supervision type, and then determine the field priority of the alarm business field based on the matched alarm business field. If the alarm message matches at least two alarm business fields, and the field priorities of the two alarm business fields are mutually exclusive, select the alarm business field with the highest field priority. Finally, based on the supervision type, alarm business field, and selected field priority of the alarm message, automatically match the supervision group, configure the supervision status, number of supervision repetitions, supervision repetition time interval, supervision delay time, and supervision area range; The data assembly unit assembles data based on the above matching information and configuration information in combination with specific alarm information, generates a supervision task, and puts it into a supervision waiting queue.

9. According to claim 8, a fault supervision system for dealing with complex scenarios based on the flink framework is characterized in that: The supervision types include alarms, dispatch orders and work orders, and the supervision waiting queues include alarm supervision waiting queues, dispatch supervision waiting queues and work order supervision waiting queues; the data assembly unit puts the corresponding generated supervision tasks into the alarm supervision waiting queues, dispatch supervision waiting queues or work order supervision waiting queues based on the supervision type of the alarm message; The supervision status includes eight types: not supervised, waiting for supervision, in supervision, successful supervision, failed supervision, repeated supervision, resumed supervision and suppressed supervision. Among them, when the status initialization unit initializes the supervision waiting queue, the supervision task in the unsupervised state is initialized to the waiting for supervision state.

10. A fault supervision system for dealing with complex scenarios based on the flink framework according to claim 9, characterized in that: When the timing polling processing module performs timing polling on the supervision tasks in the supervision waiting queue: For the supervision tasks in the alarm supervision waiting queue, when the supervision task meets the supervision delay time, the specific supervisor will be automatically selected from the configured supervision group according to the supervision area range, and then the subsequent operations will be performed; For supervision tasks in the order supervision waiting queue or work order supervision waiting queue, the order status or work order status of the supervision task is obtained through the third-party system, and when the order status or work order status of the supervision task meets the supervision delay time, the specific supervisor is automatically selected from the configured supervision group according to the supervision area range, and then subsequent operations are performed.

Images