distsql: fix copr cache events metric

[youjiali1995]

Signed-off-by: youjiali1995 [email protected]

What problem does this PR solve?

Issue Number: close #28653
Problem Summary:

The type of copr cache events metric is wrong(histogram). It should be counter.

What is changed and how it works?

What's Changed:
Fix the metric type.

How it Works:
Before:
hit and miss are always equal even if the copr cache is disabled.

After:

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

Documentation

Release note

None

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has not been approved.

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[jackysp]

LGTM

[jackysp]

Maybe simply change

tidb/metrics/grafana/tidb.json

Line 6660 in 340abf0

"expr": "sum(rate(tidb_distsql_copr_cache_count{tidb_cluster=\"$tidb_cluster\"}[1m])) by (type)",

from count to sum will also fix it?

[youjiali1995]

Maybe simply change

tidb/metrics/grafana/tidb.json

Line 6660 in 340abf0

"expr": "sum(rate(tidb_distsql_copr_cache_count{tidb_cluster=\"$tidb_cluster\"}[1m])) by (type)",

from count to sum will also fix it?

Yes, but I think histogram doesn't make sense here.

[breezewish]

Both Counter and Histogram make sense in some ways.

By using a counter, you can know the overall hit rate for all coprocessor requests.
By using a histogram, you can know the 95% SQL statement hit rate (if use a percentile function correctly).

@lzmhhh123 PTAL

[youjiali1995]

Both Counter and Histogram make sense in some ways.

By using a counter, you can know the overall hit rate for all coprocessor requests. By using a histogram, you can know the 95% SQL statement hit rate (if use a percentile function correctly).

@lzmhhh123 PTAL

TiDB uses telemetry to record the hit rate.🤯

[ti-chi-bot]

@youjiali1995: PR needs rebase.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.