OneBusAway / watchdog

76%

main: 76%

LAST BUILD ON BRANCH fix/deduplicate-unmatched-stop-ids

branch: SELECT

CHANGE BRANCH

x

• No branch selected
• Improve-config-loading
• aaronbrethorst-patch-1
• api-status-pannel
• bugfix/handle-nil-response
• cache
• ci
• ci/182-update-go-toolchain
• dh2
• docker
• docker-ci
• docker-ci-redux
• docker-hub
• docs
• docs/watchdog-server-scoped-config
• feat/158-unattributed-vehicle-reasons
• feat/158-unattributed-vehicles-clean
• feat/add-docker-compose-observability-stack
• feat/all-dashboard-visualizations
• feat/assignment-timeout-public-members
• feat/automated-versioning
• feat/exponential-backoff
• feat/freshness-aware-metrics
• feat/group-expired-issue-assignees
• feat/gtfs-bounding-box-validation
• feat/healthcheck-endpoint
• feat/increase-test-coverage
• feat/increase-test-coverage-2
• feat/inject-version-via-build-args
• feat/issue-100-vehicle-volume
• feat/issue-105-stop-matching
• feat/issue-97-dependency-latency
• feat/issue-98-feed-expiration
• feat/issue-99-agency-consistency
• feat/metrics-dashboard-skeleton
• feat/prometheus-alerting-rules
• feat/skip-unchanged-static-feeds
• feat/speed-anomaly-metrics
• feat/static-gtfs-partial-refresh
• feat/unmatched-stop-geo-clustering
• feat/watchdog-server-info
• feature/add-sentry-error-tracking
• feature/docker-setup
• feature/handle-auth-optional-in-gtfs-realtime
• feature/matching-ratios-analysis
• feature/vehicle-report-frequency
• fix/148-config-reconciliation
• fix/149-agency-mode-blank-agency-id-bbox
• fix/150-post-141-cleanups
• fix/161-freshness-gate-panels
• fix/165-gtfs-rt-fetch-freshness
• fix/Handle-GTFS-coordinate-precision
• fix/agency-mode-blank-feed-bounding-box
• fix/agency-mode-filtering
• fix/context-cancellation-graceful-shutdown
• fix/deduplicate-unmatched-stop-ids
• fix/dockerfile
• fix/metric-type-naming
• fix/metrics
• fix/per-agency-bounding-box
• fix/per-agency-vehicle-metrics
• fix/report-invalid-servers-once
• fix/reuse-metrics-response
• fix/survive-empty-remote-config
• fix/timezone-aliases
• fix/watchdog-production-environment-label
• improve-test-coverage
• main
• makefile
• merged-temp/observer-dashboard
• metrics/latlong
• oba-metrics-api
• oom-investigation
• perf/concurrent-metrics-collection
• prod-errors
• refactor/app-service-injection
• refactor/create-cache-dir-if-not-exist
• refactor/gtfs-static-in-memory
• refactor/minimal-gtfs-models
• refactor/move-utils
• refactor/reuse-http-client-realtime-store
• refactor/support-mutliple-servers
• refactor/use-helpers
• remove-unmatched-stop-locations-metric
• security
• test-ci-workflow-branch
• test/integration
• thread-safe-config
• update/readme-dockerfile-config
• ws/metrics-docs-0920-1015

Build # 39128930209

Build Type

Pull #189

github

Committed by vishnums2k5

Commit Message

fix(metrics): deduplicate unmatched stop IDs when computing unresolved metric

When OneBusAway's /api/where/metrics.json endpoint reports unmatched stop IDs for an agency, entry.StopIDsUnmatched[agencyID] can contain duplicate stop IDs if multiple vehicles or trip updates encounter the same unmatched stop in an update cycle.

gtfs.GetStopLocationsByIDs indexes stops into a map keyed by stop ID, resolving each unique stop at most once. Previously, fetchObaAPIMetrics counted resolved stops from the returned map and calculated unresolved := len(unmatchedStopIDs) - resolved. When duplicate stop IDs were present, len(unmatchedStopIDs) was greater than the number of unique stops, causing unresolved > 0 even when all unique stops were successfully resolved against the GTFS bundle.

This resulted in false-positive oba_unmatched_stop_unresolved gauge metrics and false warning logs ("OBA unmatched stop IDs could not be resolved against the local GTFS bundle").

Deduplicate unmatchedStopIDs into unique IDs before resolving and computing the unresolved count. Add a unit test to verify duplicate unmatched stop IDs do not produce a non-zero unresolved metric.

Pull Request Pull Request #189: fix(metrics): deduplicate unmatched stop IDs when computing unresolved metric

Coverage Stats

9 of 11 new or added lines in 1 file covered. (81.82%)

4384 of 5736 relevant lines covered (76.43%)

17106.72 hits per line