Adds E2E regression cover for the fix in #13812.
The Logs → Build toggle mounts PanelFieldList with pageKey "build" in
edit
mode, which deferred its stream-list fetch until a stream was already
set.
Since ZO_QUERY_ON_STREAM_SELECTION defaults to true, the logs page
selects
nothing, so Build mounted blank and its Stream dropdown was left
enabled,
openable and permanently empty on "No options found" — no way to pick a
stream, no way out.
## Tests (Logs-SelectStar shard)
- **P0** — with no stream selected on the logs page, the Build tab's
Stream
dropdown is populated and a stream can actually be picked: no empty
state,
options render, `e2e_automate` is offered, selecting it lands, and the
field list loads that stream's schema.
- **P1** — changing Stream Type with a blank stream refetches the list:
logs → traces drops `e2e_automate` (a logs stream) and loads a real
traces
list, switching back restores it. Covers the stream_type watcher, which
is
gated on the same `initialStreamsLoaded` flag the fix arms.
`beforeEach` asserts the logs page starts with no stream selected —
without
that precondition both tests would pass against the buggy build.
## Notes
- Page-object locators are scoped to
`[data-test="logs-build-query-page"]`:
the Visualize tab mounts a second PanelEditor that stays in the DOM
under
`v-show`, so an unscoped `index-dropdown-stream` is a strict-mode
violation.
- Verified 6/6 green over 3 repeats locally. The shard runs the backend
with
ZO_QUICK_MODE_ENABLED=true, which the local instance does not —
assertions
avoid named fields so quick mode's field narrowing can't affect them.
## Why
The scheduled **Alpha1 (Cloud)** e2e run failed on three jobs
([o2-enterprise run
31547517666](https://github.com/openobserve/o2-enterprise/actions/runs/31547517666)):
**Logs-Features** (7), **Functions** (11), **RegressionSet** (4).
Root-caused each against a freshly-deployed alpha; two are genuine test
heals, one was pure environment skew.
## Root causes & fixes
### 1. Logs-Features — shareLink + monaco-query-prefill (7 tests) ✅
fixed
On cloud the app builds the share link from its **public** base URL
(`*.external.zinclabs.dev`), while the tests authenticate against the
ingress host (`*.internal.zinclabs.dev`). `page.goto(sharedUrl)`
navigated **cross-origin**, dropped the auth cookies, and bounced to Dex
— so every post-redirect state assertion read back `null`.
**Fix:** re-point the short URL's origin to `ZO_BASE_URL` in
`clickShareLinkAndGetUrl()` (right next to the existing HTTP→HTTPS
"maintain auth context" step). `/short/<id>` resolves identically on the
backend regardless of ingress host, so only the origin changes and the
session stays valid. Same-origin envs (localhost / matching deployed
host) are untouched.
### 2. RegressionSet — alerts Bug #9967 PromQL (1 test) ✅ fixed
`ingestMetricsData()` authenticated with `ZO_ROOT_USER_EMAIL:PASSWORD`,
but **cloud ingestion needs `email:passcode`**. The ingest silently
401'd → the `e2e_test_cpu_usage` metrics stream was never created → the
PromQL alert flow timed out selecting it.
**Fix:** build auth via `getAuthHeaders()` (passcode on cloud, root
creds on self-hosted), refresh the passcode from the live browser
session first, and poll until the stream is indexed so the wizard's
stream-select doesn't race WAL lag.
### 3. Functions — radio → segmented toggle (11 tests) — no change
needed
The page object was already migrated to the segmented toggle in #13766;
these failures were the **alpha frontend lagging the merged test code**.
Green after the redeploy with zero code change.
## Verification (against alpha, `auto_org_1`)
- `shareLink.spec.js` + `monaco-query-prefill.spec.js` → **11/11
passed**
- `alerts-regression.spec.js` (full) → **5/5 passed** (#9967, #10110,
#10899, #10872, #10472)
- `Functions/js-transform-type.spec.js` → **7/7**,
`row-expansion.spec.js` → green
## Scope
OSS-only (test files). The ENT `playwright_alpha1.yml` checks out
opensource at `main`, so merging makes the next scheduled run pick this
up. It can also be pre-verified by dispatching that workflow with
`o2_opensource_repo=test/alpha1-e2e-heals-0812`.
Supersedes #13738, whose commit is included here as the first of four.
Adds PromQL support to count and time-sliced SLOs, lets metrics streams
choose SQL or PromQL, and fixes three shipped bugs found along the way.
## Shipped bug: every SQL time-slice SLO recorded wrong history
`fetch_rows`' `SliQueryPlan::Single` arm normalized every hit with
`with_good_column: true`, but that arm serves **two** shapes and only
one projects `zo_slo_good`:
| shape | projects | result |
|---|---|---|
| count single-query | `zo_slo_good` + `zo_slo_value` | correct |
| time slice | `zo_slo_value` only | `good` defaulted to `0.0` |
So every SQL time-slice slice was classified from `0.0` instead of its
aggregate. With a positive threshold, `<`/`<=` (e.g. `p95 < 300ms`)
recorded every slice **good** — the SLO read ~100%, the budget never
burned, and no alert on it could fire. `>`/`>=` recorded every slice
**bad** and fired continuously. Shipped in v0.92.0 (`4772df1723`);
`to_row` had no time-slice test, which is how it got through.
Fixed by deriving the flag from the SLI type in a new
`single_query_rows`. Deliberately **not** fixed inside `to_row` by
falling back to `total` — that would read a count bucket whose `SUM`
came back NULL as *fully good*, inventing uptime. A test forbids it.
**Stored history is not rewritten.** It ages out within one window; an
operator can force a rebuild sooner by making a real edit (a no-op
re-save does nothing — `update` bumps only on `definition_changed`). No
auto-migration: bumping every time-slice SLO at boot would freeze every
attached alert org-wide until backfill cleared the coverage floor, and
double each org's row budget for 97 days — a self-inflicted monitoring
gap to repair a reporting error.
## PromQL support
**Time slice** carried a `query_language` field the planner ignored, so
a PromQL time-slice SLO saved cleanly and measured nothing, forever.
- `SliQueryPlan::PromQlValue` — one expression, not a good/total pair,
planned at slice ends from the same helper the count arm uses.
- The expression is passed through **verbatim**. Grouping comes from
series labels; wrapping it in `sum by (...)` the way the SQL arm injects
GROUP BY would destroy the aggregate (the sum of four pods' p95 is not a
p95).
- `promql_value_rows` **rejects** two series colliding on one
`(slice_start, group_key)` via `RejectReason::AmbiguousSeries` instead
of summing. Summing is right for counters (what `promql_rows` does) and
meaningless here. Detection is unconditional, so the ungrouped case — an
unaggregated `histogram_quantile` returning one series per instance — is
covered. Rejects are plumbed through `fetch_rows` into
`PassResult.rejected` on both the live and backfill paths.
**Count** already had a working `CountSource::PromQl` backend that no UI
could reach.
## Metrics streams choose SQL or PromQL
`language_suits_stream` refused SQL whenever `stream_type == "metrics"`,
and the form derived the language from the stream type — so metrics had
no SQL option. A metrics stream is an ordinary stream with a `value`
column, and `SELECT histogram(_timestamp, '5 minute'), avg(value) FROM
cpu_usage` is a perfectly good SLI. SQL is now valid for any stream
type; PromQL stays metrics-only. The form asks via a toggle (defaulting
to PromQL) instead of deriving.
## Validation
- `scope` refused for PromQL — a SQL `WHERE` fragment with nowhere to go
in a PromQL plan.
- Empty `query` refused for **both** languages (the SQL aggregate was
previously never validated — same silent-freeze shape).
- Expressions parsed with `promql-parser` at save, in `core` (`config`
has no parser).
- **Parse-only by design.** A grouping check requiring the root to be an
aggregation whose `by(...)` matches `group_by` wrongly rejects
`histogram_quantile(0.95, sum by (le, region)(...))` (function-call
root; `le` is a necessary label superset) and binary-op roots like `sum
by (region)(increase(a[5m])) - sum by (region)(increase(b[5m]))`. A
"warn instead" variant needs a warnings channel the save API doesn't
have.
## Also fixed
- **Time-slice comparator 422.** `SliConfig::TimeSlice.comparator` has
no serde default, so a dropdown the user never opened was omitted and
the save died with a bare 422 — blocking creation of *any* time-slice
SLO from the UI. Seeded to `<` only when blank, so a stored comparator
survives an edit.
- **SLO/alert/service-stream regressions** (the commit from #13738):
fractional-threshold serde round-trip, `SloAlertCondition` UI, alerts
table, service streams.
## Verification
- 487 config · 206 core · 45 infra · 418 frontend — all passing; clippy
`-D warnings`, `cargo fmt`, prettier, eslint, `vue-tsc` all clean.
- Driven through a real browser against a running server (isolated
instance, own data dir). All four SLI/language combinations create and
store correctly: count·PromQL, count·**SQL over metrics**,
time-slice·PromQL, time-slice·**SQL over metrics**. Also verified: the
toggle swaps field shapes both ways, scope shows for SQL and hides for
PromQL, PromQL-only hints appear only in PromQL mode, and editing a
stored SQL-over-metrics SLO reopens as SQL with its predicate intact.
- Validation rejections confirmed live with actionable messages:
PromQL-with-scope, unparseable PromQL, and an empty SQL aggregate.
Known limitation, pre-existing: `threshold` likewise has no serde
default and is not seeded, so leaving it blank still 422s. That is
genuine user error rather than an untouched-control trap; adding form
validation was out of scope.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## Summary
Completes the Sentinel **Page Object Model** compliance sweep of the
Playwright suite (follow-up to #13727, which covered the non-Dashboards
shards). Every raw `page.locator(...)` / `page.getBy*(...)` selector —
in **actions and assertions** — is relocated out of the spec files into
the relevant page object; specs now call page-object methods/getters
only.
**Result: 0 raw fixture-`page` selectors across the entire
`playwright-tests` suite** (single-line *and* multi-line, verified with
a dotall/multiline scanner). Together with #13727, the whole suite is
POM-clean.
## What's in this PR
**Dashboards shards — ~1,460 selectors** relocated into their page
objects (`dashboard-panel-configs/actions/edit/time`, `dashboard-chart`,
`dashboard-variables-scoped`, `dashboard-filter`, `dashboard-drilldown`,
`crossLinkPage`, `visualise`, `dashboard-share-export`,
`dashboard-create`, `dashboard-settings`, `dashboard-variables`):
| Shard | selectors |
|-------|----------:|
| Dashboard-Config-Settings (11 specs) | ~255 |
| Dashboards-Variables (12 specs, 4 batches) | ~693 |
| Dashboards-Tables (7 specs) | ~127 |
| Dashboards-Charts (9 specs) | ~81 |
| Dashboards-Core + Visualize (10 specs) | ~144 |
| Dashboards-Settings + Panel-DateTime (8 specs) | ~107 |
| Dashboard-Regression (1 spec) | ~52 |
**Correlation UI specs** (3 new `Correlation/*.ui.spec.js` that arrived
via main's merge, after the original audit) — converted into a new
`correlationDrawerPage` + an extended `correlationSettingsPage`, with
`PageManager` wired into specs that previously used none.
**Audit stragglers** the original **single-line** audit grep
undercounted — multi-line `page\n.locator(...)` chains in
`logstable.spec.js` (→ `logsPage.getTableHeaderByField()`) and
`dashboard-streaming.spec.js` (2 add-panel selectors).
**AI-review fixes** (DeepSeek, on this PR) — removed 2 dead-code
duplicate getters + 1 redundant method alias (merge artifacts from the
sequential batch commits).
One Node-context `console.log` (`dashboard-general-setting`'s
`beforeEach`) → `testLogger.info`.
## Approach
Nearly every Dashboards spec shares the same page objects, so the work
was done in **sequential batches** (one committed, verified unit at a
time) to keep page-object edits conflict-free. 13 commits, mapping 1:1
to those batches.
## Not changed (by design)
- **`waitForTimeout`** left as-is — logged in the audit, not rewritten
(per direction).
- **Browser-context code inside `page.evaluate()` /
`waitForFunction()`** (`document.querySelector`, `window.*`) left as-is
— it runs in the page's JS engine, not Node, so `testLogger` (a Node
module) can't be called there. *No `console.*` remained in dashboards
specs*; the Node-context ones were converted to `testLogger`.
- **`newPage.locator(...)`** on separate popup `Page` instances left
as-is (not the fixture `page`).
## Verification
- **0 raw fixture-`page` selectors** across all specs — single- and
multi-line, comments stripped.
- All changed `.js` files pass `node --check`.
- **CI-gated** for browser execution (no local runs), consistent with
#13727. The enterprise E2E crosscheck already passed against this PR.
## Stats
82 files changed (+3,355 / −2,211) — 61 specs, 21 page objects.
Selectors **moved into page objects, not deleted**.
## What & why
Sentinel (Council of Agents) audit-driven **Page Object Model
compliance** across the Playwright suite. Many specs (a lot added
recently, some AI-generated) had raw `page.locator(...)` /
`page.getBy*(...)` selectors directly in spec files — including inside
`expect(...)` — which the Sentinel flags as CRITICAL. This PR relocates
those selectors into their page objects and calls page-object
methods/getters instead.
A full audit of **all 256 specs** was generated first (stored locally
under `tests/ui-testing/MD_Files/sentinel-audit-2026-08-10/`, which is
gitignored). Original totals: **1,609** raw selectors, **190** inside
assertions, plus console/testLogger items.
## Scope of THIS PR (non-Dashboards)
Every non-Dashboards spec is now **100% raw-selector clean** (verified:
0 occurrences). Shards covered:
- **GeneralTests, Logs (Core/Builder-Advanced/Features/Regression),
Traces (+Reg), Streams (+Reg), Metrics (+Reg), Pipelines, Reports
(+Reg), Anomaly, Login/auth, auth-warm, and the unmatched Alerts
specs.**
- `console.*` → `testLogger` where it was Node-context (kept `console.*`
**inside `page.evaluate()`** — that's browser context; `testLogger`
doesn't exist there).
- `testLogger` import/use added where missing.
- **`waitForTimeout` intentionally left as-is** (logged in the audit,
not rewritten — per direction).
## Deferred to a follow-up PR
The **heavy Dashboards shards** (Variables 693, Config-Settings 255,
Tables 127, Charts 81, Core 79, Visualize 65, Settings 52,
Panel-DateTime 55, Dashboard-Regression 52 raw selectors) are **not** in
this PR — they're the bulk of the work and will be a separate, focused
pass. (dashboard-streaming console cleanup is included here.)
## Verification
Fixes were applied and executed **shard by shard** against a local env,
gated green before commit where the env allowed:
| Group | Local result |
|-------|--------------|
| GeneralTests (edited specs) | 52 passed / 11 skipped / 0 failed |
| Logs (shareLink, logsQuickPick, indexquery) | 21 passed |
| Traces / Streams / Metrics | 49 passed / 2 skipped / 0 failed |
| Pipelines | 56 passed / 17 skipped / 0 failed |
| logs-bugs :548, dashboard-streaming | verified not regressed
(isolation) |
**Alerts create-wizard specs** (Alerts shard, Alerts-Regression,
unmatched Alerts) could not be gated locally — the alert-creation
wizard's stream popover times out on this env **on unmodified code too**
(reproduced on original, test `:363`), so those refactors are 1:1
selector→method mappings, syntax-clean (`node --check`), and rely on CI
to gate.
Every changed `.js` passes `node --check`. `waitForTimeout` unchanged;
no behavior changes intended.
## Stats
51 files changed. Selectors moved into page objects, not deleted.
🤖 Auto-generated E2E tests for #13652 — **healed to passing** against a
live OpenObserve before this PR was opened.
- Branch `autoe2e/pr-13652` (stacked on `main`), so the tests run
against the feature code and pass before it merges.
- Spec:
`tests/ui-testing/playwright-tests/Logs/streamSchemaTimezone.spec.js`
The repo Playwright CI re-validates here. Review is for **quality +
coverage** — the tests already pass.
_Generated by the E2E Council pipeline._
---------
Co-authored-by: Shrinath-O2 <214651540+Shrinath-O2@users.noreply.github.com>
## Problem
On the **alpha1 (cloud)** e2e run, the **Logs-Core** shard failed and
did not recover across all 3 retries. The single hard failure was:
`Logs/unflattened.spec.js:220 › Unflattened testcases › stream to toggle
store original data toggle and display o2 id`
### Root cause
After switching to the **unflattened** tab, the test closed the
log-detail drawer with a bare `closeDialog.waitFor()` +
`closeDialog.click()`. Switching that tab re-renders the drawer, so the
close-button node **detaches mid-click** — Playwright logs `element was
detached from the DOM, retrying` and the click hangs for the full 45s,
failing every retry identically. It's a deterministic race, which is why
retries never helped. The secondary retry-failure signature (JSON detail
panel lagging behind tab activation) is the same drawer instability,
wider on the slower shared alpha env.
## Fix
- Close the drawer via the file's existing resilient
`closeLogDetailDrawerIfOpen()` helper (Escape fallback +
wait-for-hidden) in **both** the failing test and its identical
quick-mode sibling (`unflattened.spec.js` lines ~283 and ~426).
- Harden `openJsonDetailTab()` (`pages/logsPages/unflattened.js`) to
re-assert the JSON tab and re-wait once when the panel content lags,
then fall back to the original throwing wait so a genuine non-render
still fails loudly.
Test-only change; no product code touched.
## Verification
Seeded the **Run E2E Tests on Alpha1 (Cloud)** workflow against this
branch (`o2_opensource_repo=test/alpha1-unflatten-drawer-close-flake`,
enterprise on `main`): [run
31067277452](https://github.com/openobserve/o2-enterprise/actions/runs/31067277452)
— Logs-Core and the overall run passed green.
## Summary
Adds **10 E2E tests** covering the two defects fixed in #13244, plus the
`SELECT *` behaviour change that shipped alongside them. Also introduces
per-shard backend configuration so the `SELECT *` tests can actually run
in CI.
## Test coverage
### `logsVisualizePersistence.spec.js` — 7 tests (quick mode **off**)
| # | Test | Defect | Fix code exercised |
|---|---|---|---|
| 1 | Renders without the false required-fields error | A |
`convertPanelData.ts` `!query?.customQuery` |
| 2 | Restores chart type, query and chart after URL refresh (h-bar) | B
| `setupLogsTab` re-run + `isInitialVisualizeRestore` |
| 3 | Re-renders after reload with the default chart type | B |
permanent No-Data half |
| 4 | Restores a non-default chart type (area) after refresh | B |
`restoreVisualizationFromUrlOnLoad()` |
| 5 | Chart still renders after reload with a VRL function applied | B |
`prepareVisualizeMode()` VRL copy |
| 6 | Never shows required-fields error toggling Visualize → Logs →
Visualize | A | manual-toggle rebuild path |
| 7 | No `SELECT *` toast opening Visualize with no stream in the URL |
B | page-load ordering guard |
**Defect A** — false "Please select required fields to render the chart"
on the custom-query Timechart panel.
**Defect B** — URL refresh produced a `SELECT *` toast plus permanent No
Data.
### `logsVisualizeSelectStar.spec.js` — 3 tests (quick mode **on**,
dedicated shard)
| # | Test | Behaviour verified | Call site |
|---|---|---|---|
| 8 | Table chart on `SELECT *` → toast shown **and** chart type reverts
| Table is blocked | `Index.vue:2265` (watcher) |
| 9 | Histogram chart on `SELECT *` → no toast, renders |
Line/bar/area/scatter allowed | entry path |
| 10 | `SELECT *` line chart survives browser refresh → no toast, state
restored | Restore path doesn't re-emit | `Index.vue:2465` (restore) |
## Infrastructure changes
| File | Change | Why |
|---|---|---|
| `ci_matrix.json` | New `Logs-SelectStar` shard with
`quick_mode_enabled: "true"` | `isSelectStarForTable()` only exists when
the server runs with quick mode on |
| `playwright.yml` | Per-shard `ZO_QUICK_MODE_ENABLED` env on `Start
OpenObserve` | The flag is read once at server start — a shard is the
smallest unit that can vary it |
| `build-ci-matrix.js` | Emit `quick_mode_enabled` (defaults `"false"`)
| The field was being silently stripped by the emit whitelist |
| `logsPage.js` | 13 assertion helpers + fixed chart-type detection |
Detection read `data-selected` off the wrong element and matched retired
CSS classes, so it always returned `null` |
All 24 existing shards are unaffected — they omit the field and default
to `false`.
## Notes on test quality
- Every negative assertion ("no toast", "no error") is sequenced after a
deterministic positive end-state, and gated on a precondition so it
cannot pass vacuously.
- Toast absence is proved with a MutationObserver recorder installed
before the navigation under test, since toasts auto-dismiss.
- The quick-mode precondition is a hard assertion, not `test.skip` — a
misconfigured shard fails loudly instead of silently reporting
"skipped".
## Alpha1 (Cloud) E2E — shard-by-shard failure fixes
Fixes real, root-caused failures from Alpha1 run
[30876689608](https://github.com/openobserve/o2-enterprise/actions/runs/30876689608)
(5 red shards, 16 hard failures). Each root cause was reproduced live
against alpha via Playwright MCP. **8 of 16 hard failures are fixed here
in code**; the remaining 8 are environment/data-timing driven (backend
instability during the run's 04:15–04:18 ingest window) and need a
re-run, not a code change.
### Fixes
| Shard | Fix | Root cause |
|---|---|---|
| **Alerts** (4) | `alerts-advanced.spec.js` destination URL →
`https://example.com/webhook` | The self-host URL
`${ZO_BASE_URL}/…/alerts_notify_sink/_json` resolves to a **private IP**
on cloud, so the backend **SSRF guard rejects the save** → the New
Destination form never closes → `createDestination()` times out, failing
all 4 Alerts-Advanced tests. Verified live: private-IP URL is blocked,
`example.com` (public, IANA-reserved, never delivered to) saves and
closes the form. |
| **GeneralTests** (2 + 2 flaky) | sanity `beforeEach` →
`setDateTimeToPast1Hour()` (+ new `logsPage` helper) | The
result-summary/pagination tests query the default **Past 15 Minutes**.
`e2e_automate` ingestion is bursty on shared/cloud envs — **0 rows in
15m vs 7,850 in 1h** (confirmed via API) — so the result table never
renders. This is exactly the failed-vs-flaky split seen on the run. |
| **Logs-Features** (1) | `monaco-query-prefill.spec.js` → poll
`getQueryEditorTextWhenReady()` | Lazy Monaco prefills its model
asynchronously after a saved view is applied on a fresh page load. A
one-shot read after a fixed 2s caught an empty editor (`Received: ""`).
Now polls until the query lands. Non-masking: if it never prefills, the
poll still times out and the test fails. |
| **Traces** (1 of 2) | `tracesSearch.spec.js` → poll for a settled
results/no-results state | The "no stream selected state" test read
`hasTraceResults()`/`isNoResultsVisible()` once each; the async search
settle left both momentarily false. Now polls until one terminal state
renders. |
### Not changed (deliberately)
- **Traces `traceQueryEditor.spec.js:151`** — the detail-tree check
already waits 10s; its failure means the error-trace detail didn't open,
which needs error-trace data in-window to validate. Weakening the
assertion would mask real behavior.
- **GeneralTests `enrichment.spec.js:101`, `logoManagement.spec.js:18`**
— ran during the backend-instability window and can't be reproduced
live. `[name='home']` is **not** stale (present live) and `_meta` access
is fine for the test user. Env/timing.
- **Pipelines (all 5)** — UI-click timeouts (`…-update-pipeline` /
`Explore` button not found) because pipeline creation / stream data was
unavailable while the backend returned *"No online node" / "Proxy
request" / "Expected N records, got 0"*. Same selectors 42 passing tests
in the shard use. Env-driven → re-run to confirm.
### Validation
- All fixes reproduced/verified live on alpha via Playwright MCP
(`e2etest1@o2-qa.com`, `auto_org_1`).
- All 4 edited specs parse and list (`playwright test --list`).
- Full local runs are blocked by an unrelated issue: the Dex login UI
changed (new "Continue with Email" flow) so `global-setup.js` can't
authenticate locally; CI uses the shared-auth artifact. (Worth a
separate fix.)
## Problem
In [o2-enterprise run
30791574230](https://github.com/openobserve/o2-enterprise/actions/runs/30791574230)
the **Logs-Features** shard logged **2 failed, 6 flaky, 3 skipped, 63
passed** with **52 retries**. Seven of the eight troubled tests are in
`logs-autocomplete-suggestions.spec.js` (the eighth `code`-field test
shares the cause), and every one dies on the identical assertion:
```
Error: Expected to find a field with captured values
expect(received).not.toBeNull()
Received: null
```
— i.e. the shared `findFieldWithStringValues()` helper returned `null`.
## Root cause
A **click-time race**, not missing data (the `e2e_automate` stream holds
3848 rows ingested at shard start). The helper clicked a field's sidebar
expand button and then polled IndexedDB for 10s — but the 10s window
started at **click time** and raced the `/_values_stream` fetch. In the
failing run the runner was slow (a perf assertion showed queries ~2.6×
over baseline: 1055ms vs <403ms). The values response landed **after**
the IndexedDB poll had already given up, so the record read as `null`
and the assertion failed hard. Because `findFieldWithStringValues()` is
shared across ~8 tests, one slow window knocked out the whole cluster,
and each burned up to 4 attempts → 52 retries.
An intended fix (`clickFieldExpandAndWaitValues`, a deterministic
`/_values_stream` response wait) already existed in the file but was
**never wired in**.
## Fix
- **`expandFieldAndCaptureValues()`** — anchor the IndexedDB
confirmation **after** the actual `/_values_stream` response instead of
racing it from click time; re-click to re-expand if a prior collapse
meant no fetch fired. Replaces the dead `clickFieldExpandAndWaitValues`
helper and the bare click+poll in the `code`-field test.
- **`findFieldWithStringValues()`** — scan via the anchored capture,
with a bounded refresh-and-retry round so a transient empty `/_values`
response self-heals rather than failing the test.
- **Readiness gate** — add `waitForStreamData()` to every `beforeEach`
so freshly-ingested data is searchable before the UI drives value
capture (WAL → index lag). Best-effort so an already-warm shared stream
never blocks.
## Testing
- `node --check` passes; no linter is configured for the suite.
- The slow-runner condition can't be reproduced deterministically
locally; the change is a test-harness reliability fix (deterministic
response-anchoring + readiness gate) that removes the timing dependency
that caused the flake. Will confirm green via the ENT playwright
regression this PR dispatches.
🤖 Auto-generated E2E tests for #13025 — **healed to passing** against a
live OpenObserve before this PR was opened.
- Branch `autoe2e/pr-13025` (stacked on `issue-ent-2038`), so the tests
run against the feature code and pass before it merges.
- Spec: `tests/ui-testing/playwright-tests/Logs/logsQuickPick.spec.js`
The repo Playwright CI re-validates here. Review is for **quality +
coverage** — the tests already pass.
_Generated by the E2E Council pipeline._
**Test cases added / changed:**
| Test case | Type | What it verifies |
|---|---|---|
| `should quick pick a stream from the sidebar` | new | clicking a
sidebar quick pick button selects the stream and loads the field list |
| `should show hero state on fresh arrival with no stream selected` |
new | the main content area shows both hero action cards when no stream
is selected |
| `should open stream dropdown when clicking select stream card` | new |
clicking the "Select a stream" hero card opens the stream dropdown
popover |
| `should reappear quick pick after deselecting and re-select works` |
new | after deselecting a stream the quick pick reappears and
re-selection works correctly |
| `should show more footer when org has more than 8 streams` | new | the
quick pick more footer is visible with the correct remaining count when
org has > 8 streams (skips otherwise) |
| `should select stream via hero recent chip` | new | clicking a
recent-stream chip in the hero state selects the stream and triggers
field list loading |
| `should open query guide documentation in new tab` | new | clicking
the "Read the query guide" card opens the OpenObserve docs in a new
browser tab |
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## Why
The **Logs-Features** alpha1 shard runs 18–23 min and intermittently
trips *"self-hosted runner lost communication"* — a **duration-driven
infra flake**, not a test failure (the tests were passing right up to
the runner death). A long job keeps the self-hosted runner under
sustained load, raising the odds it drops off the network. Shrink the
runtime and the flake probability drops with it.
A large part of that runtime is `waitForLoadState('networkidle')`.
`networkidle` means "wait for 500ms of zero network activity" — but the
cloud SPA is constantly chattering (telemetry / RUM / streaming search),
so that quiet moment basically never comes. Each call therefore runs to
its **full 30s ceiling**, then `.catch(() => {})` throws the result
away. Locally there's no background traffic, so the same line returns in
<1s — which is exactly why these specs fly locally but crawl on alpha.
## What
In `monaco-query-prefill.spec.js`, all **10** `networkidle` calls sit
right after a `page.goto()` (which already awaits the `load` event)
**and** right before a deterministic gate:
`selectStream`, `waitForQueryEditorContent`,
`waitForQueryEditorVisible`, or `waitForRedirectedQueryEditorContent`.
They add nothing to correctness — the following targeted wait is the
real readiness signal — so they're removed. No new timers introduced;
the existing targeted waits do the gating.
## Verified on alpha (live env, full spec)
| | Runtime | Result |
|---|---|---|
| **Before** (original) | 5.1m | 1 flaky failure (P2), last test skipped
(serial mode) |
| **After** (this change) | 4.3m | **5/5 passed** |
The after-run ran **more** tests (all 5) in **less** time, and was fully
green. A second confirmation run is in progress and will be posted
below.
## Scope
Pilot for the same cleanup. The heavy file is `searchPatterns.spec.js`
(**46** of these waits); this validates the pattern on a small,
well-understood spec first before rolling out.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## What
Stabilizes the **Logs-Queries** and **Logs-Features** shards of the
alpha1 cloud E2E
suite to green. Test-infra + page-object changes only; **no product
code** (5 files
under `tests/ui-testing/`).
> **Stacked on #13240** (Logs-Core). Base is `fix/logs-core-alpha1-e2e`
— please merge
> #13240 first; GitHub will then retarget this PR to `main`. The diff
here is only the
> Logs-Queries/Features work (5 files).
## Root causes & fixes
### Logs-Queries
**`logsDownloads.spec.js` — monolithic test exceeded the test timeout
(hard failure).**
All 14 download scenarios ran sequentially in **one** `test()` (each a
full `page.reload`
+ re-setup), needing ~7–8 min against a 360s budget. It hit 360s
mid-run, Playwright tore
the page down, and the remaining scenarios failed with **"Target
page/context closed"** (a
cascade, not a download bug). A `try/catch` wrapper also swallowed
per-scenario errors and
rethrew an aggregate, hiding the real stack. → **Split into 14 isolated
per-scenario tests.**
The `beforeEach` already yields a fresh, query-loaded page, so no
per-scenario reload is
needed; each test is short, isolated (one failing download can't close
the page for the
others), and parallelizes. Removed the error-swallowing wrapper;
`@downloadAll` tag preserved.
**`searchpartition.spec.js` — Reka OSwitch state asserted within a
hardcoded 5s (flaky).**
`verifyHistogramState` waits for the histogram OSwitch
`data-state="unchecked"` element, but
the Reka switch animates the checked→unchecked transition and the
settled state can appear
>5s after the toggle+execute under CI load. → Raised the poll ceiling to
15s (still asserts
the real end-state, no sleep).
### Logs-Features
All the share-link failures were **flaky, not product bugs** — the
redirect fully restores
every param; the tests raced the SPA under CI load.
**Reading UI/URL state before the short-URL redirect finished
hydrating.** `getQueryEditorText()`
returned `""` (lazy Monaco visible but not yet pre-filled) and
`captureCurrentState()` read
`show_histogram=null` (the SPA rewrites URL params in stages). → New
page-object helpers
`waitForQueryEditorContent`, `waitForUrlParam`, and URL-stabilization
inside
`waitForRedirectComplete` (`_waitForUrlStable`). No fixed
`waitForTimeout`.
**Bare `clickRefresh()` cancelled the in-flight SQL-mode auto-search**
("Running query
cancelled" toast), so `sql_mode`/query never committed to the short URL
→ empty editor after
redirect. → Commit via `runQueryAndWaitForResults()` (waits for the
Cancel state to clear
first), then confirm `sql_mode` is in the URL before sharing.
**SQL setup itself was unreliable.** `enableSqlModeIfNeeded` (Vue
`meta.sqlMode` flag)
didn't always stick, and `clearAndFillQueryEditor`'s focus-sensitive
`Ctrl+A` select-all
silently appended in headless CI (**doubled query**). → New
`setupSqlQueryForShare(query)`:
sets the Vue flag + writes the exact query via Monaco `executeEdits`
(verified) +
`runQueryAndWaitForResults`, then confirms `sql_mode` committed to the
URL — **retries once,
then strict-asserts** (writing SELECT text alone does **not** flip
`meta.sqlMode`; the flag
is what the app serializes to `sql_mode`). `waitForUrlParam` gained a
`required` mode that
throws instead of silently letting a bad state through to the share.
**Residual Monaco lazy-mount race after the short-URL hop** — the editor
occasionally
mounted without applying the query from the (resolved) URL. →
`waitForRedirectedQueryEditorContent`
self-heals with **one** reload of the resolved URL (the reliable
URL-param prefill path).
**Share success toast strict-mode violation** — the "Link Copied
Successfully!" toast
coexisted with the cancelled-query info toast, so the broad selector
matched 2 elements. →
Match the share-result toast by message with `.first()` in the share
helpers.
**`searchPatterns.spec.js` — async pattern extraction still
empty/timeout after 60s** under
alpha indexing load. → `setupPatternsView` re-runs the query and waits
once more before giving
up (resilience for the data dependency; callers still assert the
returned state).
## Scope / safety
- **5 files:** `pages/logsPages/logsPage.js` + specs `logsDownloads`,
`monaco-query-prefill`,
`shareLink`, `searchPatterns`. **No product code.**
- **No tests deleted, no `.only`, no unconditional skips, no assertions
removed.**
- Every fix replaces a fixed sleep / wrong assumption with a
deterministic wait on the real
condition — no `waitForTimeout`, no `force`, no retry-based masking.
## Validation
- Local on cloud alpha (`retries:0`, stricter than CI's `retries:1`):
- `logsDownloads` **14/14** (`--workers=3`) — no timeouts, no "Target
page closed".
- Previously-flaky share-link / SQL / searchpartition tests **12/12**
under `--repeat-each=3`
(`sql_mode` committed on the first attempt every time; retry never
needed).
- Full `shareLink` + `monaco-query-prefill` files **15/15** — no
regressions.
- alpha1 Logs-Queries + Logs-Features CI: green —
https://github.com/openobserve/o2-enterprise/actions/runs/29539009573🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
## What
Stabilizes the **Logs-Core** shard of the alpha1 cloud E2E suite to
green, running
**fully parallel** (`--workers=5`). Test-infra + page-object changes
only; **no product
code** (7 files under `tests/ui-testing/`).
## Root causes & fixes
### Quick mode / interesting fields (7 tests)
- **`enableQuickModeIfDisabled` treated the wrong element as the
quick-mode signal.** It
returned early whenever the field-list **user-defined-schema toggle
group** was visible —
but `FieldListPagination` renders that group whenever a stream exposes a
UDS (always true
on cloud), independent of `searchObj.meta.quickMode`. The per-field
interesting (star)
buttons are gated purely on `quickMode`, so they never rendered and
every
`ensureFieldIsInteresting` call timed out. Now verifies the **actual**
Quick Mode switch
state via the utilities menu and confirms by waiting for a per-field
interesting button.
- **`sqlModeOrderBy` hung on `waitForLoadState('networkidle')`** —
streaming keeps the
network busy so it never settles; switched to `domcontentloaded` (the
editor assertion is
the real readiness gate).
### Search-toolbar synchronization (logspage / histogram flakes)
- **`clickRefreshButton` / `clickRefresh` clicked the refresh button
mid-state-swap.** The
Run-query button swaps to a "Cancel query" variant (and briefly
detaches/disables) while a
prior/auto search is in flight; clicking then cancelled that search or
hit a
not-visible/detached node (`element is not visible`), and a subsequent
`waitForResponse('/_search')` never fired (60s timeout). Both now wait
for the run-mode
button to be **visible AND idle** (not Cancel, not disabled/`aria-busy`)
before clicking —
no `force`.
- **`clickLiveMode5Sec` + `clickRelative6WeeksButton` popovers shifted
mid-click.** The
refresh-interval dropdown and the date picker are popovers anchored to
the toolbar; while a
search is in flight the toolbar reflows and the popover repositions, so
the target never
reaches Playwright's "stable" state (~45s click timeout) or the click
misses (date range
stayed "Past 15 Minutes"). Both now wait for the toolbar to settle
(`_waitForQueryButtonIdle`) first; the date preset additionally verifies
its trigger label
and retries once.
### Histogram error-handling
- **Query-propagation race** — `clearAndFillQueryEditor` fills Monaco
but the value reaches
the store on a debounce, so `clickRefresh()` could run the **previous**
query and the
no-events/error state never appeared. Now waits for the Monaco model to
reflect the new
query (`waitForQueryEditorValue`) before running.
### Logs-table field management
- **Fixed `waitForTimeout(2000)` in `beforeEach` →
`waitForFieldListReady()`** — the
expandable field rows only render once the stream's fields have indexed
**values**, and
cloud indexing after ingestion lags under parallel load, so the sidebar
was empty
(`fieldCount === 0`, "field expand button not found"). Gates on the
data-backed field list
(60s, matching the global-setup indexing budget).
- **Severity-colors** read status bars after a fixed 3s → poll until
they render.
- **Case-insensitivity** counted matching fields after a fixed 500ms →
`expect.poll`.
- **Persist-after-refresh (app timing)** — the added column survives
reload only via the
`logFilterField` localStorage entry, written **after** the column
renders; reloading before
that commit lost it (now waits for the localStorage write). The restore
is then
intermittently slow/dropped on the first reload under cloud load, so it
reloads up to 2×
and polls for the restored column.
### jobSearch — scheduler-list hidden button (app defect, worked around)
- **App defect:** the scheduler list renders under
`v-show="showSearchScheduler"`, and the
logs page **resets that flag whenever the URL query loses the
`action=search_scheduler`
param** (`updateUrlQueryParams` strips it after a search). The "Get
Jobs" button is then in
the DOM but `display:none` and never recovers — the whole jobSearch CRUD
suite timed out.
`_getJobRowIndex` now **re-navigates to the scheduler-list URL**
(restores the param →
re-shows the list) and retries; validated firing and recovering live on
alpha. Written up
for a frontend fix (`fix/search-scheduler-job-issue`).
- **`validateAddJob` "Job added successfully" toast made best-effort** —
it's transient and
`searchSchedulerSubmit` already awaited the POST (returned the jobId),
so the toast can
auto-dismiss before the assertion runs; creation is already confirmed by
the jobId + the
scheduler-list lookup. Same re-open recovery applied to its Get-Jobs
click.
## Scope / safety
- **7 files:** `Logs` page-objects (`logsPage`, `logsQueryPage`,
`logsPageEP`,
`jobScheduler`) + Logs specs (`logshistogram`, `logsquickmode`,
`logstable`). **No product
code.**
- **No tests deleted, no `.only`, no unconditional skips, no assertions
removed.**
- Every fix replaces a fixed sleep / wrong assumption with a
deterministic wait on the real
condition — no `waitForTimeout`, no `force` in the touched paths.
## Validation
- Local on cloud alpha: `logsquickmode` 8/8, `logshistogram` 5/5,
`logspage` 23/23,
`logstable` field-management + persist (`--repeat-each` clean),
jobSearch scheduler-list
recovery observed firing and recovering live.
- alpha1 Logs-Core CI: green —
https://github.com/openobserve/o2-enterprise/actions/runs/29471282407🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Shrinath Rao <shnath@openobserve.ai>
🤖 Auto-generated E2E tests for #11813 — **healed to passing** against a
live OpenObserve before this PR was opened.
- Branch `autoe2e/pr-11813` (stacked on `main`), so the tests run
against the feature code and pass before it merges.
- Spec:
`tests/ui-testing/playwright-tests/Logs/logsVisualizePersistence.spec.js`
The repo Playwright CI re-validates here. Review is for **quality +
coverage** — the tests already pass.
_Generated by the E2E Council pipeline._
**Test cases added / changed:**
| Test case | Type | What it verifies |
|---|---|---|
| `should persist live mode / auto-run toggle across navigation` |
extend | live-mode toggle state-guard added before navigation to catch
silent no-op when auto_query_enabled is false |
---------
Co-authored-by: Shrinath-O2 <214651540+Shrinath-O2@users.noreply.github.com>
Co-authored-by: Hengfei Yang <hengfei.yang@gmail.com>
## Reduce `Logs-Builder-Basic` e2e shard runtime — ~10 min → ~5 min
### 🎯 Result (measured on CI)
| Metric | Before | After | Saved |
|--------|--------|-------|-------|
| **`Logs-Builder-Basic` shard** | **10m 05s** | **5m 12s** | **~4m 53s
(≈48% faster)** |
The shard runtime is cut roughly in half. Most of the old time was fixed
sleeps, waits that always ran their full timeout, and duplicated
per-test setup — not actual test logic. **No test cases, steps, or
assertions were removed.**
### What we changed
| # | Area | Before | After | Saving |
|---|------|--------|-------|--------|
| 1 | `initQueryBuilderTest` setup | Logs page loaded twice (helper
`goto` + `selectStream` navigates again) | Navigate once | 1 page load ×
~41 tests |
| 2 | `waitForBuildTabLoaded` | Always waited up to 10s for editor to
fill, even when it's meant to be empty | Skip when SQL mode is off |
~10s dead wait on many tests |
| 3 | `enableSqlModeIfNeeded` | 2s watcher wait + 500ms sleep, even
where the watcher never fires | Gated to Build tab; waits on real signal
| ~2.5s × ~35 tests |
| 4 | `disableSqlModeIfNeeded` | Fixed 500–600ms sleeps | Deterministic
wait on editor update | Blind sleeps removed |
| 5 | `applyQueryButton` | Clicked Run blindly | Waits for button ready
first | Fewer early-click retries |
| 6 | `pagination.spec` | Fresh stream ingest + index-wait per test (9×)
| One ingest per worker, reused | 8 redundant ingests + waits |
| 7 | `pagination.spec` | Dead `waitForTimeout(2000)` per streaming test
| Removed | ~2s × 5 tests |
| 8 | `unflattened.spec` | Search run twice with 3s + 1s sleeps | Single
run | ~5s × ~8 calls |
| 9 | `unflattened` row scan | Fixed 3s sleep per row | Wait for content
to render | 15-row scan: ~97s → seconds |
| 10 | `unflattened.spec` | Blind sleep after ingest | Poll until stream
queryable | No over-waiting |
### Bonus
Items 2–5 live in the shared `logsPage` page object, so
`Logs-Builder-Advanced`, `Logs-Core`, and `Logs-Features` shards get
faster for free too.
### Notes
| Item | Detail |
|------|--------|
| `playwright.yml` | Unchanged — shard layout stays as-is |
| Assertions | Preserved everywhere (e.g. pagination still checks the
same record counts) |
| Validation | All 52 tests still discovered by `playwright --list`;
shard confirmed green on CI at 5m 12s |
## What
Gate the full **Playwright (UI)** and **API** test suites on PRs behind
a single **`e2e` label**, and **auto-bypass** PRs that only touch files
which can't affect those suites. Covers `playwright.yml` and
`api-testing.yml`.
## Behaviour (both suites)
| Situation | build/test jobs | summary check (`playwright_summary` /
`api_tests_summary`) | Mergeable? |
|---|---|---|---|
| PR with `e2e` label | full suite runs | green if tests pass | ✅ if
green |
| PR, no label, relevant code changed | skipped | **fails** ("add `e2e`
label") | ❌ blocked |
| PR, no label, docs/meta only | skipped | **passes** (auto-bypass) | ✅
|
| Remove the `e2e` label | skipped (fresh run) | **fails** again | ❌
until re-added |
| push to `main` / merge queue / dispatch | full suite runs | green if
tests pass | n/a |
One `e2e` label runs **both** UI + API. `push`, `merge_group`, and
`workflow_dispatch` are never label-gated.
## Why the gate lives in the summary job
A job that `if:`-skips reports `skipped`, which branch protection treats
as **pass**. So a no-label PR can't be blocked by merely skipping jobs —
the summary job **actively fails** when the change is relevant and the
`e2e` label is absent.
## Auto-bypass (`check_changes`)
No label needed for: `*.md/.mdx/.markdown/.rst/.adoc`, `docs/`,
images/media, `LICENSE/NOTICE/AUTHORS/CONTRIBUTORS/CHANGELOG/.mailmap`,
`CODEOWNERS`, GitHub issue/PR/funding templates,
`dependabot`/`renovate`,
`.gitignore/.gitattributes/.editorconfig/.gitkeep/.husky/.vscode/.idea`.
The **API** workflow additionally bypasses `web/` + UI-test dirs
(frontend changes don't affect API tests).
**Not** bypassed (still trigger): `.github/workflows/**` (incl. these
workflow files), source, `package*.json`/lockfiles, configs, `*.txt`.
## Rollout prerequisites
- Create the **`e2e`** label in this repo.
- Required status checks = **`playwright_summary`** and
**`api_tests_summary`** — the individual shard jobs must NOT be required
(they legitimately skip and would auto-pass).
Sibling PR on the other repo uses the same branch `ci/e2e-label-gate`.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
🤖 Auto-generated E2E tests for #12857 — **healed to passing** against a
live OpenObserve before this PR was opened.
- Branch `autoe2e/pr-12857` (stacked on `main`), so the tests run
against the feature code and pass before it merges.
- Spec:
`tests/ui-testing/playwright-tests/Logs/ftsDefaultColumn.spec.js`
The repo Playwright CI re-validates here. Review is for **quality +
coverage** — the tests already pass.
_Generated by the E2E Council pipeline._
**Test cases added / changed:**
| Test case | Type | What it verifies |
|---|---|---|
| `should clear FTS selection on stream change and re-resolve for new
stream` | extend | switching streams clears the carried-over FTS default
selection, then re-resolves the FTS default column for the new stream on
the next search |
| `should clear FTS default on field reset and re-resolve on next
search` | extend | clicking the field-list reset icon clears the
auto-picked FTS default column from the table, then running a new search
re-resolves it |
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Shrinath-O2 <214651540+Shrinath-O2@users.noreply.github.com>
Reduces three independent e2e flakes. All test-infra only; no product
code.
## 1. Pipelines — `ingestData` "Premature close" (the original failure)
`LogsPage.ingestData` failed all records with `Failed to ingest N out of
N`. Server returned HTTP 200 and ingested every record, but node-fetch
2.7.0 reused a keep-alive socket the server closed during the 500ms
inter-record gap → "Premature close" reading the body (0/384 succeeded).
Fix: disable keep-alive (fresh connection per request, per-protocol
agent) + per-record retry. Verified: Pipelines shard green.
## 2. Dashboards-Core — `dashboard-chartJson.spec.js` popup flake
Custom-query tests (1 failed + 3 flaky) failed at the JSON/non-timestamp
checkbox with "popper intercepts pointer events" / detached timeouts.
Root cause (reproduced 100% locally): the field-property ODropdown
auto-closes within ~300ms when opened right after the panel's first
render — a late field-list re-render dismisses it. `toBeVisible()`
passes at t=0 but the checkbox is gone by click time. Fix:
`openFieldPropertyPopup` now opens, confirms the popup STAYS open, and
re-opens if it auto-closed (throws if it never stays open — no masking).
Verified locally across repeated runs.
## 3. Logs-Builder — `unflattened.spec.js` `_o2_id` readiness
Failed with `Failed to find _o2_id field after 3 attempts`; CI log shows
the UI detail was empty (`fields: []`) — re-ingested data wasn't
search-visible yet on the contended runner. The readiness gate existed
but its result was ignored, so the test fell through to a doomed UI scan
and burned all 4 retries (~26 min). Fix: respect the gate's result (on
timeout, wait once more and log), and widen the UI row scan 10→15. The
final assertion is unchanged — a genuinely absent `_o2_id` still throws.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## Summary
- Add E2E test coverage for SQL queries containing `--` inline comments
in the Logs search page
- Fix verified: queries like `-- comment\nSELECT _timestamp FROM stream`
no longer produce `SQL ParserError("Expected: end of statement, found:
_timestamp")`
- Add two reusable page object methods to `logsPage.js` for asserting
SQL error state
## What Was Broken
SQL queries with `--` style comments followed by a `SELECT` statement
were failing with:
> `SQL error: ParserError("Expected: end of statement, found: _timestamp
at Line: 1, Column: 45")`
The parser was not stripping comment lines before executing the query,
causing it to misread the `SELECT` column list as a second statement.
## Test Scenarios Covered
| # | Scenario | Tag |
|---|----------|-----|
| 1 | Single-line comment before SELECT | `@P0 @smoke` |
| 2 | Multiple comment lines before SELECT | `@P1` |
| 3 | Inline comment at end of a line (`SELECT ... -- comment`) | `@P1`
|
| 4 | Comment in the middle of a multi-line SQL query | `@P1` |
| 5 | Comment + WHERE clause using `str_match_ignore_case` | `@P1` |
| 6 | Comment-only query — must NOT produce `_timestamp` parser error
(edge case) | `@P2` |
## Files Changed
-
`tests/ui-testing/playwright-tests/Logs/logsqueries.sql-comments.spec.js`
← new spec file
- `tests/ui-testing/pages/logsPages/logsPage.js` ← two new assertion
methods added
- `.github/workflows/playwright.yml` ← spec registered in CI pipeline
---------
Co-authored-by: Abhay Padamani <141122205+ktx-abhay@users.noreply.github.com>
## Problem
The **Logs-Builder-Basic** e2e shard is frequently flaky — it times out
on `unflattened.spec.js` and, because it's a big shard, a single timeout
re-runs the whole thing. Example: [run
27521591530](https://github.com/openobserve/openobserve/actions/runs/27521591530/job/81341218030).
In that run:
- `unflattened.spec.js:205` — hard failure, **timed out (300s) on all 3
retries**
- `unflattened.spec.js:100` and `logsQueryBuilder-chart.spec.js:148` —
flaky (passed on retry)
## Root cause
The failures are **timeouts, not assertion bugs**. The unflattened
tests:
1. Slept **~76s of fixed `waitForTimeout`** across the file (pure dead
time).
2. Ran a retry loop that **re-ingested 3,848 records up to 5×/test**
while waiting for the `_o2_id` field to index after enabling *Store
Original Data*.
Each test + setup is ~80s on a fast local machine; CI runners are ~3-4×
slower **and** run all 4 specs in the shard in parallel against one
backend. That pushes individual tests right onto the 300s/test ceiling.
The re-ingestion balloon also hammered the shared backend, slowing the
other specs (`logsQueryBuilder`, `pagination`) — which is why they
flaked as collateral.
## Fix
- Add `waitForO2IdQueryable()` to the unflattened page object: **polls
the search API until `_o2_id` is actually queryable** instead of
sleeping a fixed interval and hoping indexing caught up.
- Gate the UI scan on that readiness signal, removing the 15s/5s blind
schema-propagation sleeps.
- Turn the expensive **re-ingest** retry loop into a cheap **re-query**
loop (attempts 5 → 3) — the gate guarantees the data is already in the
backend, so a UI miss is only render lag.
No change to what the tests actually assert.
## Verification (local, against self-hosted OpenObserve)
| Check | Result |
|---|---|
| Unflattened solo ×2 (cleanup between) | 2 passed each, `_o2_id` on
**row 0 / attempt 1** every time, **2.3m** (was 2.8m) |
| `logsQueryBuilder-chart:148` isolated ×5 | **5/5 passed** → not
independently flaky (collateral) |
| **Full 4-spec shard, 4 workers (CI scenario)** | **52 passed, 0
failed, 0 flaky** |
The retry loop now never fires — the readiness gate makes the first UI
scan succeed every run, killing both the timeout and the cross-spec
contention.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## Summary
- Adds a single GeneralTests spec (`edition-features.spec.js`) that
validates the header edition button, the `EnterpriseUpgradeDialog`,
every enterprise feature card, and the `o2.ws/*` documentation links
each card opens.
- Two tests live in the spec; exactly one runs per environment. The
active edition is detected from the rendered header button label (most
direct signal of what the user sees, tolerates frontend/backend
build-flag mismatches like an OSS frontend served by an ENT backend).
- Pairs with the same-named branch on `openobserve/o2-enterprise` so ENT
CI tree-merges this spec in and runs the ENT variant.
## What the spec validates
- Header button label per edition (`Get OpenObserve Enterprise (Free)`
vs `Edition: Enterprise`)
- Hero panel (title, badge, primary button, features title, Learn More
button)
- 20 enterprise feature cards (presence + external-link icon + BETA/HA
badges)
- Cloud-only `Bring Your Own Bucket` card stays hidden outside Cloud
- Every feature card click opens a popup that lands on the expected
`openobserve.ai` doc page
- Hero external link buttons (`Download Now`/`Learn More` on OSS, `Learn
More` on ENT) — `Get Free License` intentionally not clicked because it
switches the active org to the meta org
## Files
-
`tests/ui-testing/playwright-tests/GeneralTests/edition-features.spec.js`
(new)
- `tests/ui-testing/pages/generalPages/editionFeaturesPage.js` (new —
page object + truth table of slugs / expected doc paths)
- `tests/ui-testing/pages/page-manager.js` (register
`pm.editionFeaturesPage`)
## Test plan
- [x] OSS local run: 1 passed, 1 skipped (~2.0m)
- [x] ENT local run: 1 passed, 1 skipped (~2.0m)
- [x] Sentinel audit passed (POM compliance, no raw selectors in spec,
no hardcoded secrets, no console.log)
- [ ] OSS CI green on this PR
- [ ] ENT CI green on companion PR
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
## Summary
- **13 Playwright E2E journey tests** for the Model Pricing feature
(\`/settings/model_pricing\`), consolidating all P0–P2 scenarios into
journey-based flows
- New page object \`ModelPricingPage\` under \`pages/generalPages/\`
- API cleanup integrated into the centralized \`APICleanup\` class and
called from \`cleanup.spec.js\`
- CI matrix updated: \`ZO_MODEL_PRICING_ENABLED=true\` env var added;
\`model-pricing.spec.js\` added to the \`GeneralTests\` job
- **Bug fix**: \`waitForBuildTabLoaded\` now polls the Monaco editor for
non-empty content (Phase 3), fixing a timing race where Build tab SQL
tests read an empty query string on slower CI machines
## Test coverage
| Journey | Scenarios covered |
|---------|------------------|
| List, tabs & search | Controls, tab switching, search, built-in tab |
| Create & validation | All form validation cases + successful save |
| Edit, cancel & delete | Full lifecycle |
| Duplicate & clone | Org model duplicate + built-in clone |
| Bulk actions | Select, export JSON, bulk delete |
| Toggle | Disable/re-enable with correct visual state |
| Test match dialog | Empty state, no-result, clear, match result |
| Import | Valid JSON, correction flow, cancel |
| Editor features | Multi-tier, pricing template, pattern examples |
| Pricing drawer | Overflow link → drawer with all price keys |
| Empty state | CTA when org has no custom models |
| Cost applied | LLM span matched → cost stored correctly |
| No matching model | Unknown model name → zero cost stored |
## Bug fix: Build tab query editor timing
\`waitForBuildTabLoaded\`'s Phase 2 (chart-type-selected signal) fires
as soon as \`initializeFromQuery\` sets the chart type — before
\`makeAutoSQLQuery\` has run. On slower CI machines this left
\`queries[0].query\` empty when tests read the editor, causing the "SQL
ON + Auto mode" and "Auto → Custom" tests to receive \`""\` instead of
generated SQL.
Phase 3 polls the Monaco editor
(\`[data-test="logs-search-bar-query-editor"]\`) for non-empty content,
confirming \`onBuildQueryGenerated\` has been called and the query is
ready.
## Test plan
- [x] All 13 model pricing tests pass locally
- [x] \`logsQueryBuilder-editor.spec.js\` (23 tests) — all pass with
Phase 3 fix applied
- [x] Sentinel code audit passed (no blocking issues)
- [x] API cleanup added to \`cleanup.spec.js\` via
\`apiCleanup.cleanupModelPricingModels()\` — 404-safe for instances
where the feature is disabled
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
## Summary
Two independent improvements to the Playwright E2E test suite.
---
### Part 1 — Sentinel audit fixes (original PR scope)
Fixes from the Sentinel audit review of PR #12358 (selector drift
stabilization).
- **Dead code removal**: delete `clickJobID()` from `logsPageEP.js`
(unreachable code containing `console.log`); remove unused `chromium`
import
- **Critical fix**: replace `console.log`/`console.error` in
`decryptLogSQL()` with `testLogger` equivalents
- **POM compliance**: replace raw `page.locator()` call in
`alerts-ui-operations.spec.js` `beforeEach` with
`pm.alertsPage.isIncidentsFeatureEnabled()`; add that method to
`alertsPage.js`
- **POM compliance**: move per-test `const pm = new PageManager(page)`
into `beforeEach` in `service-graph.spec.js`; replace raw
`page.waitForLoadState()` with
`pm.serviceGraphPage.waitForGraphReload()`; add `waitForGraphReload()`
to `serviceGraphPage.js`
- **Inline selector cleanup**: move metrics tab inline string literals
in `serviceGraphPage.js` into constructor properties
- **Tagging**: add
`@cipherKeys/@enterprise/@P1/@P2/@negative/@validation` to all 6 tests
in `cipherKeys.spec.js` (were untagged); add `@P1/@P2` to all 5 tests in
`regexPatternManagement.spec.js`
- **ESM fix**: correct `export const path = require('path')` → `import
path from 'path'` in `logoManagementPage.js`
- **Deduplication**: consolidate
`createFirstFunction`/`createSecondFunction` into shared
`createFunction(name, vrlCode)` in `pipelinesEP.js`
---
### Part 2 — Parallelize Metrics E2E shard
**Problem**: The `e2e / Metrics` CI shard was consistently the slowest
job at ~22 minutes per PR run, while all other shards complete in ~5-7
minutes.
**Root cause**: 9 of 11 Metrics spec files used
`test.describe.configure({ mode: 'serial' })` because they shared a
describe-scope `let pm` variable reassigned in `beforeEach`. This forced
every test in those files onto a single Playwright worker, eliminating
all parallelism despite `fullyParallel: true` and 5 workers being
configured.
**Fix applied to all 9 affected files** (`metrics.spec.js`,
`metrics-config.spec.js`, `metrics-config-tabs.spec.js`,
`metrics-visualizations.spec.js`,
`metrics-promql-query-persistence.spec.js`,
`metrics-table-column-order.spec.js`, `metrics-queries.spec.js`,
`metrics-aggregations.spec.js`, `metrics-advanced.spec.js`):
- Remove `test.describe.configure({ mode: 'serial' })`
- Replace shared `let pm` + `beforeEach` pattern with a `setupTest(page,
testInfo)` helper that creates and returns a fresh `PageManager` per
test — the same pattern already used in `metrics-promql-builder.spec.js`
(the one file that was already parallel)
- `metrics-table-column-order.spec.js`: pass `pm` explicitly into
`setupTableChart(pm, page, mode)` since it can no longer close over a
describe-scope variable
- Split consolidated mega-tests (for-loops over multiple independent
scenarios) in `metrics-queries`, `metrics-aggregations`, and
`metrics-advanced` into individual `test()` blocks so Playwright can
distribute them across workers
**No test assertions, selectors, query logic, or validation changed.**
| File | Tests before | Tests after |
|---|---|---|
| `metrics.spec.js` | 10 (serial) | 10 (parallel) |
| `metrics-config.spec.js` | 2 (serial) | 2 (parallel) |
| `metrics-config-tabs.spec.js` | 3 (serial) | 3 (parallel) |
| `metrics-visualizations.spec.js` | 4 (serial) | 4 (parallel) |
| `metrics-promql-query-persistence.spec.js` | 6 (serial) | 6 (parallel)
|
| `metrics-table-column-order.spec.js` | 20 (serial) | 20 (parallel) |
| `metrics-queries.spec.js` | 4 mega-tests (serial) | 12 individual
(parallel) |
| `metrics-aggregations.spec.js` | 3 mega-tests (serial) | 10 individual
(parallel) |
| `metrics-advanced.spec.js` | 4 mega-tests (serial) | 12 individual
(parallel) |
**Expected impact**: ~1300s (~22 min) → ~300-400s (~5-7 min) wall-clock
time.
---
## Test plan
- [ ] All 40 tests in the 4 Sentinel-audit spec files enumerate cleanly
via `npx playwright test --list`
- [ ] Tag filters `@cipherKeys`, `@P2`, `@sdr` match expected tests
- [ ] All Metrics spec files enumerate cleanly with no serial mode
warnings
- [ ] Metrics shard CI time drops to ~5-7 min on next PR run
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Neha P
Omkar Kesarkhane
Prabhat Sharmaandgithub-actions[bot]
Vamsi Karthik R
BhargavandHengfei Yang
ktx-akshay
