Transaction log collection method, device and system

Abstract

The invention provides a transaction log collection method, device and system. The transaction log collection method comprises the following steps: putting log files of application servers into a cache queue; determining a time difference according to the tail transaction time and the head message transaction time; determining log output time according to the time difference and the header message transaction time; and sorting the log files in the cache queue according to the transaction time in the log files, and putting the sorted log files into an output message queue according to the log output time. The method can effectively deal with a large amount of logs generated in a short time, ensures that the collected logs are complete and orderly, avoids transaction failure, and further improves the playback efficiency and playback effect of a subsequent account following playback device.

Description

technical field [0001] The present invention relates to the technical field of big data data analysis, in particular to a transaction log collection method, device and system. Background technique [0002] figure 1 is a schematic diagram of transaction log generation. Such as figure 1 As shown, the user initiates a request to the backend through various channels provided by the bank, such as opening an account, lending money, etc., and the server will record each request information (Req, that is, a request message) and response information (Rep, that is, a response message) text) to the log. The transaction playback system reproduces transactions based on logs. The log message collection system is used to collect this message information and send it to the Kafka message queue for consumption by the account playback system. [0003] The banking system is a distributed system, and the logs are distributed on each application server (AP) and updated in real time. The coll...

AI Technical Summary

Problems solved by technology

If the scanning of application server 1 has just ended at this time, it needs to wait for 1 second to start scanning. At this time, the message information of application server 2 may be recorded in the message queue in advance. Similarly, the request message of application server 3 also has It may be collected before request 1 and request 2, and the follow-up playback system will have timing confusion when playing back the received messages, resulting in transaction failure

Images