:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
feat(taxonomic-filter): measure search latency on the all tab (#117280)
posthog[bot]93737617069d
diff --git a/.agents/skills/modifying-taxonomic-filter/references/telemetry-contract.md b/.agents/skills/modifying-taxonomic-filter/references/telemetry-contract.mdindex 9a27e64a379e..1489ec8faac3 100644--- a/.agents/skills/modifying-taxonomic-filter/references/telemetry-contract.md+++ b/.agents/skills/modifying-taxonomic-filter/references/telemetry-contract.md@@ -69,7 +69,10 @@ All carry `surface` on both surfaces, except `include stale toggled`, which carr## Legacy only, with `surface`-- `taxonomic filter search latency`: `groupType`, `searchQuery`, `resultCount`, `time_to_see_data_ms`.+- `taxonomic filter search latency`: `groupType`, `searchQuery`, `resultCount`, `time_to_see_data_ms`, `revealedBy`, `pendingGroupTypes`.+ A category tab sends it when its own fetch lands, and `time_to_see_data_ms` is the fetch time.+ The All tab sends it when its rows reveal, with `groupType: suggested_filters`, `revealedBy` and `pendingGroupTypes`.+ There `time_to_see_data_ms` runs from the last keystroke, so it includes the input debounce. Compare the two only by `groupType`.The rebuild does not send it yet (see `hooks/useGroupList.ts`).- `taxonomic filter search intent predicted`: `query`, `activeTab`, `predictedGroupType`, `confidence`, `isConfident`, `suggestsSwitch`, `wouldPromote`, `promptVersion`, `source`, `variant`, `inTime`, `shown`.The banner funnel in [refreshing-product-reality](refreshing-product-reality.md) reads `variant` and `shown`.diff --git a/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.test.ts b/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.test.tsindex d806403d208b..f1004f0e480d 100644--- a/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.test.ts+++ b/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.test.ts@@ -693,9 +693,21 @@ describe('taxonomicFilterLogic', () => {barrierLogic.actions.setSearchQuery('whatever')}).toMatchValues({ revealBarrierOpen: false })+ const captureSpy = jest.spyOn(posthog, 'capture')await expectLogic(eventsListLogic).toDispatchActions(['loadRemoteItemsSuccess'])await expectLogic(barrierLogic).toDispatchActions(['openRevealBarrier']).delay(1)expect(barrierLogic.values.revealBarrierOpen).toBe(true)++ const latencyCalls = captureSpy.mock.calls.filter(([event]) => event === 'taxonomic filter search latency')+ expect(latencyCalls).toHaveLength(1)+ expect(latencyCalls[0][1]).toMatchObject({+ groupType: TaxonomicFilterGroupType.SuggestedFilters,+ searchQuery: 'whatever',+ time_to_see_data_ms: expect.any(Number),+ revealedBy: 'all_groups_settled',+ pendingGroupTypes: [],+ })+ captureSpy.mockRestore()})})diff --git a/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.tsx b/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.tsxindex 456d537a0944..da8efc10c1fc 100644--- a/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.tsx+++ b/frontend/src/lib/components/TaxonomicFilter/taxonomicFilterLogic.tsx@@ -2644,6 +2644,7 @@ export const taxonomicFilterLogic = kea<taxonomicFilterLogicType>([// 5s reveal-barrier timer as a normal keystroke would — the `setSearchQuery`// listener doesn't run on mount because no action was dispatched.if ((props.initialSearchQuery ?? '').trim()) {+ cache.searchStartedAt = performance.now()cache.disposables.add(() => {const timerId = window.setTimeout(() => actions.openRevealBarrier(), REVEAL_BARRIER_TIMEOUT_MS)return () => window.clearTimeout(timerId)@@ -2832,11 +2833,13 @@ export const taxonomicFilterLogic = kea<taxonomicFilterLogicType>([// key auto-disposes the previous timer; we only need to explicitly dispose when// the query clears (no fresh timer to schedule).if ((searchQuery ?? '').trim()) {+ cache.searchStartedAt = performance.now()cache.disposables.add(() => {const timerId = window.setTimeout(() => actions.openRevealBarrier(), REVEAL_BARRIER_TIMEOUT_MS)return () => window.clearTimeout(timerId)}, 'revealBarrierTimer')} else {+ cache.searchStartedAt = nullcache.disposables.dispose('revealBarrierTimer')}@@ -2858,6 +2861,30 @@ export const taxonomicFilterLogic = kea<taxonomicFilterLogicType>([}},+ openRevealBarrier: () => {+ // The All tab runs no fetch of its own, so `infiniteListResultsReceived` never reports its+ // latency. Its rows appear when this barrier opens, so the wait the user sees ends here.+ // A group that still loads at this point means the timeout opened the barrier.+ const startedAt = cache.searchStartedAt+ cache.searchStartedAt = null+ if (+ typeof startedAt !== 'number' ||+ values.activeTab !== TaxonomicFilterGroupType.SuggestedFilters ||+ !values.searchQuery+ ) {+ return+ }+ posthog.capture('taxonomic filter search latency', {+ surface: legacyTaxonomicSurface(),+ groupType: TaxonomicFilterGroupType.SuggestedFilters,+ searchQuery: values.searchQuery,+ time_to_see_data_ms: Math.floor(performance.now() - startedAt),+ resultCount: values.infiniteListCounts[TaxonomicFilterGroupType.SuggestedFilters],+ revealedBy: values.anyGroupLoading ? 'timeout' : 'all_groups_settled',+ pendingGroupTypes: values.loadingGroupTypes,+ })+ },+setIncludeStaleEvents: ({ includeStaleEvents }) => {posthog.capture('taxonomic filter include stale toggled', {includeStaleEvents,