Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
alexluong
commented
Mar 8, 2026
alexbouchardd
requested changes
Mar 9, 2026
Split LogStore into Records + Metrics sub-interfaces. LogStore is now the combined interface so all existing consumers are unaffected. Typed responses per endpoint (EventMetricsResponse, AttemptMetricsResponse) with all fields as pointers. Stub implementations for CH, PG, and mem drivers return errNotImplemented. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a comprehensive metrics test dataset spanning January 2000 with 300 tenant-1 events (50 sparse across 5 days + 250 dense bell-curve on Jan 15) and 5 tenant-2 events for isolation. All dimension cycling produces round numbers (100/topic, 150/dest, 180 success, 120 failed, error_rate=0.4, etc). Covers all granularities (1m, 1h, 1d, 1w, 1M), dimensions, filters, and measures for both event and attempt metrics. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Dynamic SQL builder for event and attempt metrics with parameterized queries, time bucketing (date_bin/date_trunc), dimension grouping, conditional aggregates (FILTER WHERE), 30s query timeout fallback, and row limit enforcement with truncation flag. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Naive implementation using uniqExact/uniqExactIf for dedup-safe aggregation over ReplacingMergeTree without FINAL. Includes 30s query timeout fallback and 100k row limit with truncation detection. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
All three drivers (mem, pg, ch) now implement the metrics interface. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Wire up QueryEventMetrics and QueryAttemptMetrics as GET /api/v1/metrics/events and GET /api/v1/metrics/attempts with query param parsing, allowlist validation, JWT tenant scoping, and response transformation. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…re, add metrics OpenAPI spec memlogstore QueryEventMetrics was missing tenant_id dimension causing empty results when grouping by tenant. QueryAttemptMetrics was missing both tenant_id and attempt_number dimensions. Also adds MetricsResponse schemas and /metrics/events, /metrics/attempts paths to the OpenAPI spec. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Split metrics drivertest into DataCorrectness (existing value assertions) and Characteristics (structural contract tests for dense bucket filling, ordering, alignment, deterministic count, zero measures, no-data ranges, and dimension × time filling). Shared dataset setup, single provisioning. Characteristics tests will fail until bucket filling is implemented. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Enables new(expr) syntax for cleaner pointer initialization. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Sparse time-series responses caused dashboard charts to render fat bars instead of slim bars across all time slots. This adds a shared bucket filling layer (internal/logstore/bucket/) called by all 3 backends after query, producing dense responses with zero-filled gaps. - Extract TruncateTime into shared bucket package - FillEventBuckets / FillAttemptBuckets with dimension-aware filling - Update drivertest assertions for dense bucket counts - All characteristics tests now pass Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The API allowlist accepted these filters but all three query builders (pg, ch, mem) silently ignored them, causing filters like attempt_number=0 to return unfiltered results. Add WHERE clauses in all drivers and conformance tests to prevent regression. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ateRange to TimeRange in Go code Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Merge main (PR #732) and update metrics handlers to use ParseArrayQueryParam and resolveTenantIDsFilter. Update OpenAPI spec filters to oneOf string/array schema with bracket notation. Update test query strings to indexed bracket format. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add per-second throughput rate measures computed as count / bucket_duration_seconds. Events endpoint gets `rate`, attempts gets `rate`, `successful_rate`, `failed_rate`. Rate computation lives in shared driver/rate.go, called by each driver after bucket filling. Dependency measures are auto-enriched (e.g. requesting `rate` without `count` internally adds `count` for SQL but omits it from API response). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
alexbouchardd
approved these changes
Mar 11, 2026
6-chart metrics dashboard: delivery volume (stacked bar), error rate (line, 0-100%), retries (multi-line), avg attempt number, status code breakdown, and topic breakdown. Shared timeframe selector (1h/24h/7d/30d), all charts use /metrics/attempts endpoint. Includes dataviz CSS vars for info (blue) and warning (orange) themes. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…hart Restructure the metrics grid to 3 rows: events/deliveries, error breakdown (3-col), and retry pressure. Add new "Events / count" chart using attempt_number=0 filter. Support title/subtitle pattern with muted subtitles and add filters param to useMetrics hook. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add sparkline + event count to each destination row using 4h granularity with attempt_number=0 filter. Includes Sparkline component with stacked success/failed bars, empty-bar rendering, and granularity override for useMetrics hook. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Move Loading import to top of MetricsChart.tsx. Use arrays/objects directly in useMetrics instead of serializing to strings and re-splitting. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- seed_metrics.sh: generates realistic event→attempt chains with configurable error rates, retry chains, and time distribution - qa_metrics.sh: 11 named scenarios (healthy, failing, spike, empty, single, all-fail, all-success, recent, many-topics, many-codes, retry-heavy) with verification checklists - README documenting usage and scenarios Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Collaborator
Author
|
Added seed & QA scripts in
Ran full manual QA against Postgres. All 11 scenarios passing across all timeframes (1h/24h/7d/30d). Dashboard renders correctly: error rate Y-axis shows percentages, stacked bars work, breakdowns sort correctly, empty states render, single-event edge case works, |
Add manual=false filter alongside attempt_number=0 to prevent manual retries (which also start at attempt_number=0) from inflating event counts in the destination metrics chart and destinations list sparkline. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Manual retries start a new chain at attempt_number=0, inflating first_attempt_count. Add AND NOT manual to CH, PG, and memlogstore queries. Add FIXME for test dataset which assigns manual and attempt_number independently. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…anularity, filters Migrate bench suite to current metrics API (TimeRange, tenant via filters) and add bench cases for rate measures, multi-value granularities (2d/w/M), new dimensions (code, attempt_number), and new filters (code, manual, attempt_number, multi-filter). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
PR #740 changed attempt_number from 0-based to 1-based. Update all metrics query logic, test data, seed scripts, and bench seeds to match. Co-Authored-By: Claude Opus 4.6 <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
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.
implements #210