Dashboard Agent: investigations, watches, alerts, and notifications - #4418

Draft

kathiekiwi wants to merge 168 commits into

mainfrom

feat/dashboard-agent-flows

Commits on Jul 13, 2026

feat(metrics-pipeline): generic Redis-stream to ClickHouse metrics pipeline
ericallam
committed
feat(clickhouse): queue metrics tables and read queries
ericallam
committed
feat(run-engine): emit queue depth, throughput, and scheduling-delay signals
ericallam
committed
feat(webapp): queue metrics ingestion, admin controls, and emission switch
ericallam
committed
feat(tsql): opt-in gap-fill for time-bucketed series
ericallam
committed
feat(webapp): Queues dashboard and per-org metrics UI flag
ericallam
committed
chore(webapp): add server-changes note for queue metrics
ericallam
committed
chore: apply oxfmt formatting
ericallam
committed
chore: use import type for type-only imports
ericallam
committed
fix(tsql): avoid polynomial backtracking in ORDER BY direction strip
ericallam
committed
fix(tsql): strip ORDER BY direction without a backtracking regex
ericallam
committed
fix(clickhouse): remove semicolons from queue metrics migration comments
ericallam
committed
test(clickhouse): rewrite queue metrics test for cumulative counters
ericallam
committed
fix(metrics-pipeline): use BigInt order keys and namespaced odometer counters
ericallam
committed
fix(clickhouse): filter zero waits from quantile view and accept string order keys
ericallam
committed
fix(webapp): fail open on queue metrics and honor sparkline total override
ericallam
committed
test(run-engine): import describe from vitest in run-queue metrics test
ericallam
committed
fix(tsql): skip gap-fill on descending bucket order
ericallam
committed
fix(metrics-pipeline): widen order_key packing factor to 1e6
ericallam
committed
feat(webapp,clickhouse): standard time filter for queue metrics pages
ericallam
committed
fix(tsql): register the deltaSumTimestampMerge aggregate
ericallam
committed
chore(webapp): use shared primitives on the admin queue metrics page
ericallam
committed
feat(webapp): house style hero charts on the queues list
ericallam
committed
feat(clickhouse): queue activity ranking queries
ericallam
committed
feat(webapp): queue allocation view and relevance-ordered queue list
ericallam
committed
fix(tsql): inject time fallbacks into FROM subqueries
ericallam
committed
feat(clickhouse,tsql): queue metrics rollups and single-scan ranking
ericallam
committed
feat(webapp): serve queue metrics reads from rollups and fix env totals
ericallam
committed
feat(clickhouse,webapp): keep 10-second resolution on the env metrics rollup
ericallam
committed
fix(webapp): include rollup tables in the queue metrics simulator reset
ericallam
committed
fix(webapp): update the queue metrics simulator for cumulative counters
ericallam
committed
feat(webapp): stage fake Redis usage from the queue metrics simulator
ericallam
committed
fix(webapp): keep the search filter on the ranked queue list's tail
ericallam
committed
fix(metrics-pipeline): drop metric emits while the metrics Redis is not ready
ericallam
committed
feat(webapp): remove auto-balance from the allocation view
ericallam
committed
feat(tsql,webapp): reject cross-queue merges of per-queue counter states
ericallam
committed
refactor(clickhouse,webapp): use the stored quantile list in all merge calls
ericallam
committed
fix(tsql): satisfy the tenant column type in the merge guard test schema
ericallam
committed
fix(clickhouse): dedup retried metric batches in the MV target tiers
ericallam
committed
fix(webapp,metrics-pipeline): review hardening for overflow counters, cluster slots, and stream caps
ericallam
committed
feat(run-engine,clickhouse,webapp): concurrency-key health metrics and live per-key breakdown
ericallam
committed
feat(webapp,run-engine,clickhouse,tsql): per-concurrency-key queue metrics history
ericallam
committed
feat(webapp): move concurrency-key metrics to a tab on the queue detail page
ericallam
committed
feat(webapp,tsql): hide the per-key metrics table from user-facing query surfaces
ericallam
committed
fix(run-engine): anchor scheduling delay to the current queue stint on re-enqueue
ericallam
committed
feat(webapp): surface queue metrics on task detail and run inspector
ericallam
committed
chore(webapp): remove trailing blank line flagged by oxfmt
ericallam
committed
refactor(webapp): flatten the run inspector queue block and tab the task detail queue chart
ericallam
committed
fix(clickhouse): renumber queue-metrics migration to 036 to avoid a version collision
ericallam
committed

Commits on Jul 21, 2026

Commits on Jul 22, 2026

Commits on Jul 24, 2026

Queue metrics & health: UI polish (TRI-12068) (#4354)
samejr
authored

Commits on Jul 27, 2026

Commits on Jul 28, 2026

Commits on Jul 29, 2026

Commits on Jul 30, 2026