Occasional performance regression on benchmarksql
- Dominant language
- Go
- Stars
- 40.5k
- Forks
- 6.2k
- PR merge metrics
- PR metrics pending
Description
## Bug Report
There is occasional performance regression on benchmarksql.
releated issue: #55426
### 1. Minimal reproduce step (Required)
1. Deploy a cluster with 3 TiDB and 3 Tikv.
2. Run benchmark benchmarksql (benchmarksql-1k) multiple times until find the specific io or cpu usage pattern on grafana.
### 2. What did you expect to see? (Required)
No performance regression.
QPS: 64039 (+6.4% compared to baseline 60193)
or
QPS: 61259 (+1.77% compared to baseline 60193)
### 3. What did you see instead (Required)
Performance regression .
QPS: 55187 (-8.3% compared to baseline 60193)
or
QPS: 54639 (-9.20% compared to baseline 60193)
### 4. What is your TiDB version? (Required)
pingcap/tidb@26443dab89ba5b1a866de3c3e44e2f51468bc435
tikv/tikv@7c612787d0bfc57a339907ebcb26777dc296fd5f
tikv/pd@098b802fcda75b5ae0527b0b50b95ba8b31403cf
or
pingcap/tidb@74034d4ac243b3c14dbf5f8a9edb92e740da4212
tikv/tikv@de06b38a118582ee01d16ab4d3582bd8fa5cb635
tikv/pd@76bc34aee22554353bd7ea9785d3772dcb7ab31b
Contributor guide
Research direction
Start by reviewing related issue #55426 and reproducing benchmarksql-1k repeatedly on the specified 3-TiDB/3-TiKV cluster. Compare the Grafana CPU or IO patterns and QPS across runs and the listed TiDB, TiKV, and PD revisions; done means the cause of the regression is identified and the benchmark no longer shows the reported drop.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go, grafana, sql
- Domain
- databases, distributed-systems, observability, performance
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100