feat(repo): schema encode slow metrics + entity state - #862
Merged
Conversation
Widen encode/decode duration histogram buckets into multi-second stalls, count app.schema.slow (≥100ms) with entity labels, and annotate spans with app.schema.slow plus app.entity.state on encode. Supports Honeycomb p99/rate alerts on heavy aggregates (e.g. BauhausOrder) that block the event loop. Co-authored-by: Patrick Roza <42661+patroza@users.noreply.github.com>
@effect-app/cli
effect-app
@effect-app/eslint-codegen-model
@effect-app/eslint-shared-config
@effect-app/infra
@effect-app/vue
@effect-app/vue-components
commit: |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
Prod Honeycomb (scanner APIs) shows event-loop delay max spikes of ~0.5–1.5s that line up with repository schema encode, not PDF work. The worst entity is BauhausOrder (max encode ~1.0s); MultiPickDeliveryNote / MakoOrder also have multi-hundred-ms tails. Existing metrics already label
app.entity, but:What
app.schema.{encode,decode}.durationhistogram buckets through multi-second stallsapp.schema.slowwhen duration ≥ 100ms, withapp.entity+ operation attributesapp.schema.slow, and on encodeapp.entity.statefrom the first item’s_tag(so we can see which Bauhaus order states burn CPU)Dig notes (BauhausOrder encode path)
Why a single-item save can still cost 0.5–1s:
Repository.saveAndPublish— one Order document embeds nesteditems[](tagged line-item phases) +packages[](tagged package/BB phases) +ReadonlyMappacks, addresses, etc. Cosmos body ~30KB avg / ~80KB max.OrderState× item/package unions) — encode walks the full tree even for tiny field changes.app.schema.encode.duration_mson the save span has no child spans for the CPU portion.encodeKeys→ id only; resolver is decode-path, not the encode stall.After this lands + scanner bumps: alert on
RATE(app.schema.slow)grouped byapp.entity, and filter spans withapp.schema.slow=true/app.entity.state.How
timeSchemainpackages/effect-app/src/Model/Repository/internal/internal.ts— histogram bounds, slow counter, span attrs; encode call sites pass state tag.Remarks
SUM/RATE app.schema.slowwhereapp.entity in (BauhausOrder, MakoOrder, MultiPickDeliveryNote)— not overall encode p99.Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is disabled.