Query Plan too large may cause query SLOW_QUERY faild
Open
severity/moderate
sig/diagnosis
type/bug
type/enhancement
- Dominant language
- Go
- Stars
- 40.5k
- Forks
- 6.2k
- PR merge metrics
- PR metrics pending
Description
## Problem
Since slow log store the original plan and doesn't trim even the plan is very large, then the user query `SLOW_QUERY` or `CLUSTER_SLOW_QUERY` maybe failed cause by the query memory usage more than `mem-quota-query`, then the query will be killed by TiDB.
As I known, the biggest query plan after decodePlan, the size is more than `4GB`, already bigger than the default value of `mem-quota-query` `1GB`
Contributor guide
Assessment
This issue has not been assessed yet.