Support replaying audit logs #866
Description
Feature Request
Traffic files are generated by TiProxy but many clusters don't have TiProxy deployed, so we can support replaying from audit logs.
- Support reading from TiDB audit logs #859
- Deduplicate audit logs and ignore uninitialized statements #862
- Support reading audit log files in order #864
- Support replaying prepared statements #867
- Update prepared statement ID in the replay command #870
- Support replaying with a specified start time #874
- Support reading multiple traffic paths concurrently #875
- Replay in the client mode #876
- Support ignoring errors in the replayer #889
- Decode audit logs in advance #888
- List S3 file with prefix #892
- Add logs for replay progress #904
- Support
--ps-close=nevermode #922 - Reading S3 and decoding commands in parallel #940
- Add more fields in the report tables #929
- Support graceful shutdown of the replayer #927
- Support dry-run mode in traffic replay #938
- Record and load the timestamp from disk for replay #950
- Filter the commands by command end time #959
- Init DB when replay instead of parsing #964
- Automatically detect new TiDB audit log path #968
- Output replayed SQL into log files #974
- Support start replayer in service mode #970
- Add a new argument
wait-untilto start the replay job later #1001 - Support no auto quit mode for replayer #1004
- Connect to itself in the dry-run mode #969
- Output the heap profile when the memory usage is high #979
- Limit the decoded commands within a memory usage
- Filter retried audit logs by the RETRY field #1008
- Support comparing execution time and affected rows from audit logs and group by SQL digest #1014