Deprecate vttablet metrics QueryCacheXX and rename to QueryPlanCacheXX

[frouioui]

Description

This PR is a follow up of #14947, where we added two new metrics but in the end we would like to rename to TabletQueryPlanCache for more clarity and to avoid the ambiguity with an actual query cache.

All the QueryCacheXX metrics in VTTablet are renamed.

We will need to update the release notes of v20.0.0 and backport only the release notes change to release-20.0, I am adding the backport to label, but the backport should not be merge without changing its content.

Before

After

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[frouioui]

This flag was removed a while ago, I took the liberty of removing this from the config file.

[codecov]

Codecov Report

Attention: Patch coverage is 61.53846% with 10 lines in your changes missing coverage. Please review.

Project coverage is 68.68%. Comparing base (75aee8a) to head (ade9210).
Report is 7 commits behind head on main.

Files	Patch %	Lines
go/vt/vttablet/tabletserver/query_engine.go	66.66%	8 Missing ⚠️
go/vt/vttablet/tabletserver/debugenv.go	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16289      +/-   ##
==========================================
- Coverage   68.68%   68.68%   -0.01%     
==========================================
  Files        1548     1548              
  Lines      199081   199103      +22     
==========================================
  Hits       136747   136747              
- Misses      62334    62356      +22     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[deepthi]

Can you please address the naming issue?
For metrics PRs, we should always provide a screenshot showing the changed section from the output of the /metrics endpoint and /debug/vars endpoint. Looking at /metrics would have immediately shown you what I said in my review comment.

[deepthi]

The reason metric names are not prefixed with Tablet or Gate is that the binary name is automatically added to the metric name as a prefix. So this becomes vttablet_query_plan_cache_misses if you look at /metrics.
All of the metrics introduced in this PR need to be renamed and the Tablet prefix should be removed.

[deepthi]

Maybe you just need to drop the last commit!

[frouioui]

That won't work as the vtgate and vttablet metrics will have the same name in vtcombo, leading to a panic/failure.

[frouioui]

I will have to see how I can tweak vtcombo. We might already have this situation where the same name is used by both vtgate and vttablet in vtcombo.

[deepthi]

Why not do a separate PR just for the v20 release notes? then this can take as much time as necessary.

[frouioui]

Why not do a separate PR just for the v20 release notes? then this can take as much time as necessary.

That's a good point: #16299

[deepthi]

In order to ensure that we keep testing the old metrics until we delete them, we should keep the tests for them. That will require keeping the struct members also around until the metrics are deleted. Each vttablet has one instance of the queryEngine, so it should not be a big deal.

[frouioui]

Done via ade9210

[deepthi]

We should do a similar PR for vtgate after this is complete.

[frouioui]

We should do a similar PR for vtgate after this is complete.

VTGate is already using the proper name.

vitess/go/vt/vtgate/executor.go

Lines 195 to 212 in 75aee8a

	stats.NewGaugeFunc("QueryPlanCacheLength", "Query plan cache length", func() int64 {
	return int64(e.plans.Len())
	})
	stats.NewGaugeFunc("QueryPlanCacheSize", "Query plan cache size", func() int64 {
	return int64(e.plans.UsedCapacity())
	})
	stats.NewGaugeFunc("QueryPlanCacheCapacity", "Query plan cache capacity", func() int64 {
	return int64(e.plans.MaxCapacity())
	})
	stats.NewCounterFunc("QueryPlanCacheEvictions", "Query plan cache evictions", func() int64 {
	return e.plans.Metrics.Evicted()
	})
	stats.NewCounterFunc("QueryPlanCacheHits", "Query plan cache hits", func() int64 {
	return e.plans.Metrics.Hits()
	})
	stats.NewCounterFunc("QueryPlanCacheMisses", "Query plan cache misses", func() int64 {
	return e.plans.Metrics.Hits()
	})