Execute both Availability Group reads against a real Postgres (#991) - #1697
Merged
Conversation
The two AG store reads shipped in #1692 with no test that runs their SQL. That is the one gap that mattered, because the defect they were fixed for is invisible to a unit test: the grains were briefly read as two statements over a SINGLE command to save a round trip, and PostgreSQL rejects that. Npgsql only splits multi-statement text into a batch when it parses the SQL for NAMED placeholders, so with the positional ($1) parameters these reads use it sends one extended-protocol Parse and the server answers "cannot insert multiple commands into a prepared statement". Because every AG path is failure-isolated, that would not have crashed anything. It would have logged one error per server per sweep with the entire alert family silently dead - the worst failure mode a monitoring product has, since the thing that is broken is the thing that tells you something is broken. It was caught in review rather than by the suite. Added to the existing DARLING_TEST_PG-gated live section (skipped locally, run by CI's "Darling PostgreSQL tests" job), seeding both collector tables and asserting the things only real SQL can prove: that each query executes at all, that the MAX(collection_time) predicate keeps an older snapshot out, that a row with a NULL identity column is dropped rather than keyed under a placeholder, that NULL lag and suspend-reason columns round-trip, and that the whole thing reaches EvaluateStoreAlertsAsync and fires exactly one alert off freshly seeded rows. The cleanup helper issues one command per statement for the same reason. Darling suite green: 3250 passed, 156 skipped (the new test skips without a store; it is the +1 against the previous 155). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
erikdarlingdata
enabled auto-merge
July 26, 2026 18:44
pull Bot
pushed a commit
to ehtick/PerformanceMonitor
that referenced
this pull request
Jul 29, 2026
Concurrency groups for build.yml and sql-validation.yml, keyed per PR number for pull_request events only. Push and release runs get a unique per-run group (run_id), so dev/main integration builds, release builds, and every push commit's required-check result are never queued behind, replaced by, or cancelled through this mechanism. Rationale: the shared Windows runner pool is what serializes everyone's CI (erikdarlingdata#1697 sat queued behind two dev builds), so a stale PR run cancelled at re-push time is reclaimed capacity for whatever sits queued behind it. claude-review.yml already had exactly this shape; build.yml and sql-validation.yml now match it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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
1 participant
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.
Follow-up to #1692. The two Availability Group store reads shipped there with no test that runs their SQL, and that was the one gap that mattered.
Why this is not routine coverage
The defect those reads were fixed for is invisible to a unit test. Both grains were briefly read as two statements over a single command to save a round trip. PostgreSQL rejects that: Npgsql only splits multi-statement text into a batch when it parses the SQL for named placeholders, so with the positional (
$1) parameters every read in that file uses, it sends one extended-protocolParseand the server answerscannot insert multiple commands into a prepared statement.Because every AG path is failure-isolated, it would not have crashed anything. It would have logged one error per server per sweep with the entire alert family silently dead -- the worst failure mode a monitoring product has, since the thing that is broken is the thing that tells you something is broken. It was caught in review, not by the suite. Only running the SQL catches that shape.
What it asserts
Seeds both collector tables and checks the things only real SQL can prove:
MAX(collection_time)predicate keeps an older snapshot outEvaluateStoreAlertsAsyncand fires exactly one alert off freshly seeded rows (the disconnect and the suspend are first sightings, so they are silent baselines; the 900-second lag is past the 300-second default)Added to the existing
DARLING_TEST_PG-gated live section, so it skips locally and runs in CI's Darling PostgreSQL tests job. The cleanup helper issues one command per statement for the same reason as the code under test.Testing
Darling suite green: 3250 passed, 156 skipped -- the new test is the +1 against the previous 155 skipped, i.e. it is correctly gated. It executes for real in the CI Postgres job on this PR.
🤖 Generated with Claude Code