fix(infra): switch eval alarm to simple metric (no SEARCH)#141
Merged
Conversation
PR 4c's setup_eval_quality_alarm.sh used a SEARCH expression to reduce across (judged_agent_id, criterion, judge_model) combos at alarm-evaluation time. Caught at deploy: ValidationError: SEARCH is not supported on Metric Alarms. CloudWatch dashboards support SEARCH; alarms don't. Switched to a simple metric alarm against the new agent_quality_score_4w_mean_min floor metric that the rolling-mean Lambda emits (alpha-engine-research PR #96). --namespace AlphaEngine/Eval --metric-name agent_quality_score_4w_mean_min --statistic Minimum --period 86400 --evaluation-periods 1 --threshold 3.0 --comparison-operator LessThanThreshold Operator workflow unchanged: alarm fires → operator clicks dashboard quality-trend page to identify which combo triggered. The SEARCH design wouldn't have surfaced per-combo identity in the alarm body either. bash -n syntax-clean. Tests 433 still pass (no test logic affected). Deploy after merge: ./infrastructure/setup_eval_quality_alarm.sh (idempotent — overwrites the alarm in place). Co-Authored-By: Claude Opus 4.7 (1M context) <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Caught at deploy — PR 4c's alarm script failed with `ValidationError: SEARCH is not supported on Metric Alarms`. CloudWatch dashboards support SEARCH; alarms don't. Switched to a simple metric alarm against the new `agent_quality_score_4w_mean_min` floor metric (paired with research PR #96).
```
--namespace AlphaEngine/Eval
--metric-name agent_quality_score_4w_mean_min
--statistic Minimum --period 86400 --evaluation-periods 1
--threshold 3.0 --comparison-operator LessThanThreshold
```
Operator workflow unchanged
Alarm fires → operator clicks dashboard quality-trend page to identify which combo triggered. The SEARCH design wouldn't have surfaced per-combo identity in the alarm body anyway.
Pairs with
`alpha-engine-research` PR #96 — emits the floor metric this alarm reads.
Deploy after merge
`./infrastructure/setup_eval_quality_alarm.sh` (idempotent, overwrites in place).
Test plan
🤖 Generated with Claude Code