metrics(flatkv): add FlatKV observability metrics and logs#3366
Open
blindchaser wants to merge 4 commits intomainfrom
Open
metrics(flatkv): add FlatKV observability metrics and logs#3366blindchaser wants to merge 4 commits intomainfrom
blindchaser wants to merge 4 commits intomainfrom
Conversation
|
The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).
|
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #3366 +/- ##
==========================================
+ Coverage 59.07% 59.20% +0.12%
==========================================
Files 2099 2098 -1
Lines 172988 172801 -187
==========================================
+ Hits 102195 102307 +112
+ Misses 61922 61625 -297
+ Partials 8871 8869 -2
Flags with carried forward coverage won't be shown. Click here to find out more.
🚀 New features to boost your workflow:
|
cody-littley
approved these changes
May 5, 2026
Comment on lines
+440
to
+449
| otelMetrics.SnapshotWriteLatency.Record(s.ctx, secondsSince(start), | ||
| metric.WithAttributes(successAttr(err))) | ||
| if err == nil { | ||
| otelMetrics.CurrentSnapshotHeight.Record(s.ctx, s.committedVersion) | ||
| } else if !errors.Is(err, errReadOnly) { | ||
| logger.Error("FlatKV snapshot failed", | ||
| "version", s.committedVersion, | ||
| "elapsed", time.Since(start), | ||
| "err", err) | ||
| } |
Contributor
There was a problem hiding this comment.
Consider encapsulating this logic into a helper function inside flatkv/metrics.go. This is already fairly complex and long code, and so separating the metrics+logging stuff into helper functions might be worth it. (This is an optional suggestion, feel free to push back or ignore if you disagree.)
Comment on lines
+578
to
+589
| defer func() { | ||
| otelMetrics.RollbackLatency.Record(s.ctx, secondsSince(start), | ||
| metric.WithAttributes(successAttr(err))) | ||
| if err == nil { | ||
| otelMetrics.CurrentVersion.Record(s.ctx, s.committedVersion) | ||
| } else if !errors.Is(err, errReadOnly) { | ||
| logger.Error("FlatKV Rollback failed", | ||
| "targetVersion", targetVersion, | ||
| "elapsed", time.Since(start), | ||
| "err", err) | ||
| } | ||
| }() |
Contributor
There was a problem hiding this comment.
Similar to above, consider making this a helper function. Optional.
Comment on lines
+270
to
+293
| defer func() { | ||
| otelMetrics.OpenLatency.Record(s.ctx, secondsSince(start), | ||
| metric.WithAttributes(attribute.Bool("read_only", readOnly), successAttr(retErr))) | ||
| if retErr == nil { | ||
| version := s.committedVersion | ||
| if opened != nil { | ||
| version = opened.Version() | ||
| } | ||
| if !readOnly { | ||
| otelMetrics.CurrentVersion.Record(s.ctx, s.committedVersion) | ||
| } | ||
| logger.Info("FlatKV LoadVersion complete", | ||
| "targetVersion", targetVersion, | ||
| "readOnly", readOnly, | ||
| "version", version, | ||
| "elapsed", time.Since(start)) | ||
| } else if !errors.Is(retErr, errReadOnly) { | ||
| logger.Error("FlatKV LoadVersion failed", | ||
| "targetVersion", targetVersion, | ||
| "readOnly", readOnly, | ||
| "elapsed", time.Since(start), | ||
| "err", retErr) | ||
| } | ||
| }() |
Contributor
There was a problem hiding this comment.
Optional suggestion, make this a helper function
Comment on lines
+19
to
+30
| func (s *CommitStore) ApplyChangeSets(changeSets []*proto.NamedChangeSet) (err error) { | ||
| start := time.Now() | ||
| defer func() { | ||
| otelMetrics.ApplyChangesetsLatency.Record(s.ctx, secondsSince(start), | ||
| metric.WithAttributes(successAttr(err))) | ||
| if err != nil && !errors.Is(err, errReadOnly) { | ||
| logger.Error("FlatKV ApplyChangeSets failed", | ||
| "changesets", len(changeSets), | ||
| "elapsed", time.Since(start), | ||
| "err", err) | ||
| } | ||
| }() |
Contributor
There was a problem hiding this comment.
Optional, make this a helper function
Comment on lines
+96
to
+112
| defer func() { | ||
| otelMetrics.CatchupLatency.Record(s.ctx, secondsSince(start), | ||
| metric.WithAttributes(successAttr(err))) | ||
| if replayed > 0 { | ||
| otelMetrics.CatchupReplayNumBlocks.Add(s.ctx, int64(replayed)) | ||
| otelMetrics.CurrentVersion.Record(s.ctx, s.committedVersion) | ||
| } | ||
| if err != nil { | ||
| logger.Error("FlatKV catchup failed", | ||
| "targetVersion", targetVersion, | ||
| "startOffset", startOff, | ||
| "endOffset", endOff, | ||
| "replayed", replayed, | ||
| "elapsed", time.Since(start), | ||
| "err", err) | ||
| } | ||
| }() |
Contributor
There was a problem hiding this comment.
optional, make this a helper function
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR adds a FlatKV observability layer around the existing persistence pipeline. It centralizes OpenTelemetry latency histograms, counters, and gauges while preserving the WAL -> per-DB batch -> metadata commit ordering and crash recovery model.
sei-db/state_db/sc/flatkv/metrics.go: defines the FlatKV meter, shared instruments, boolsuccessattributes, db attributes, duration helpers, pending-write gauges, version/snapshot gauges, and KV pair counters.sei-db/state_db/sc/flatkv/store.go: propagates the top-levelEnablePebbleMetricsflag into each nested PebbleDB config, records writerLoadVersionlatency/current version, and initializes snapshot-height gauges on open. Readonly clones do not overwrite writer version gauges.sei-db/state_db/sc/flatkv/store_apply.go: recordsApplyChangeSetslatency, applied KV counts, pending-write gauges, and debug fields that distinguish raw changes from merged writes. LtHash updates remain after classification and old-value reads.sei-db/state_db/sc/flatkv/store_write.go: records commit latency, per-DB batch commit latency, flush latency, pending-write resets, and current version. Commit error paths return the attempted version so deferred failure logs report the version being committed.sei-db/state_db/sc/flatkv/store_catchup.go: records catchup latency, replayed block counts, current version, and progress/failure logs while retaining WAL replay and global metadata commit behavior.sei-db/state_db/sc/flatkv/snapshot.go: records snapshot write latency, prune latency, current snapshot height, prune attempts, and rollback latency while preserving atomic snapshot swap and rollback ordering. Readonly guard failures are not logged as operational errors.sei-db/state_db/sc/flatkv/importer.go: records import latency, worker flush latency, imported KV counts, import stats, and post-import version/snapshot gauges while preserving per-DB worker ownership and final snapshot durability.sei-db/state_db/sc/flatkv/store_test.go: assertsInitializeDataDirectoriesenforces the top-level Pebble metrics setting across all nested DB configs.Test plan
sei-db/state_db/sc/flatkv/store_test.go:TestInitializeDataDirectoriesPropagatesPebbleMetricscovers propagation ofEnablePebbleMetrics=falseto account, code, storage, legacy, and metadata PebbleDB configs.