Dashboard show txn_start_ts value is different from tidb_slow_query.log

[yilongrong]

Bug Report

Please answer these questions before submitting your issue. Thanks!

What did you do?

1. When I find tikv.log include slow-query information, I need use txn_start_ts value to find the sql.

[tracker.rs:199] [slow-query] [internal_key_skipped_count=970284] [internal_delete_skipped_count=0] [block_cache_hit_count=657] [block_read_count=0] [block_read_byte=0] [encrypt_data_nanos=0] [decrypt_data_nanos=0]
[scan_first_range="Some(start: 7480000000000000715F728000000000277903 end: 7480000000000000715F72800000000036472F)"]
[scan_ranges=1] [scan_iter_processed=970284] [scan_iter_ops=970286] [scan_is_desc=false] [tag=select] [table_id=113]
[txn_start_ts=419447227966554113]
[handler_build_time=0ns] [snapshot_wait_time=1ms]
[schedule_wait_time=0ns] [wait_time=1ms] [total_process_time=2.501s]
[req_time=2.641s] [remote_host=ipv4:172.16.5.169:34238] [region_id=2113]

2. Dashboard show txn_start_ts value is not correct.

What did you expect to see?

Dashboard show txn_start_ts value should same with slow query log.

What did you see instead?

What version of TiDB Dashboard are you using (./tidb-dashboard --version)?

v 4.0.5

[baurine]

Thank you for your report, let me have a check.

[baurine]

It seems not easy to reproduce. hi @yilongrong , can you make sure the above two are the exact same slow query? For example the same time, the same SQL...

[breezewish]

I believe this issue can be fixed by #793