mirror of
https://github.com/openobserve/openobserve.git
synced 2026-11-04 08:09:27 +00:00
1224 lines
62 KiB
YAML
1224 lines
62 KiB
YAML
name: Playwright UI Tests
|
|
|
|
on:
|
|
# No `push: main` trigger: the merge_group (merge-queue) run already builds and
|
|
# tests the exact commit that lands on main, so a post-merge push run would be a
|
|
# full-suite duplicate. The merge_group run is now the trusted main-ward build and
|
|
# also warms the shared Rust cache (see save-if on the rust-cache step below).
|
|
pull_request:
|
|
branches:
|
|
- "**"
|
|
# labeled/unlabeled so 'ready-for-ci' gates CI; only a new commit cancels an in-flight run (see concurrency); a draft starts no paid CI (an in-flight run still finishes).
|
|
# auto_merge_enabled: the one full-matrix run per PR, at merge time (see generate_matrix).
|
|
types: [opened, reopened, synchronize, labeled, unlabeled, auto_merge_enabled, ready_for_review, converted_to_draft]
|
|
merge_group:
|
|
schedule:
|
|
- cron: '30 20 * * *' # 2:00 AM IST
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.event.merge_group.head_ref || github.ref }}
|
|
# Only a new commit cancels: a cancelled same-SHA run leaves a failed required summary that a newer run cannot clear.
|
|
cancel-in-progress: ${{ github.event_name != 'pull_request' || github.event.action == 'synchronize' }}
|
|
|
|
env:
|
|
COLUMNS: 150
|
|
ZO_ROOT_USER_EMAIL: root@example.com
|
|
ZO_ROOT_USER_PASSWORD: Complexpass#123
|
|
ZO_BASE_URL: http://localhost:5080
|
|
WS_ZO_BASE_URL: ws://localhost:5080
|
|
ZO_BASE_URL_SC: http://localhost:5080
|
|
ZO_BASE_URL_SC_UI: http://localhost:5080
|
|
INGESTION_URL: http://localhost:5080
|
|
ORGNAME: default
|
|
ZO_WEB_URL: "http://localhost:5080"
|
|
ZO_QUICK_MODE_NUM_FIELDS: 100
|
|
ZO_QUICK_MODE_STRATEGY: first
|
|
ZO_ALLOW_USER_DEFINED_SCHEMAS: true
|
|
ZO_INGEST_ALLOWED_UPTO: 5
|
|
ZO_SLO_BACKFILL_CHUNK_SECS: 86400 # matches the server default; per-shard override below
|
|
ZO_FEATURE_QUERY_EXCLUDE_ALL: false
|
|
ZO_USAGE_BATCH_SIZE: 200
|
|
ZO_USAGE_PUBLISH_INTERVAL: 2
|
|
ZO_MIN_AUTO_REFRESH_INTERVAL: 5
|
|
ZO_STREAMING_ENABLED: true
|
|
ZO_COLS_PER_RECORD_LIMIT: "80000"
|
|
# SMTP was already enabled here, but with no host/port it fell back to the
|
|
# server defaults (localhost:25) where nothing listens — so destinations could
|
|
# be SAVED (the enabled-gate passes) while every send failed silently. The
|
|
# Mailpit step below gives that flag something to talk to, which also makes the
|
|
# delivery assertions in Alerts/dl-email-destinations.spec.js runnable.
|
|
ZO_SMTP_ENABLED: true
|
|
ZO_SMTP_HOST: localhost
|
|
# 1026 is the DL relay, not Mailpit — it expands DL_ADDRESS and forwards
|
|
# everything else through untouched, so non-email shards see no difference.
|
|
ZO_SMTP_PORT: 1026
|
|
ZO_SMTP_FROM_EMAIL: alerts@ci.openobserve.test
|
|
# Read by tests/ui-testing/playwright-tests/utils/mail-sink.js
|
|
MAILPIT_URL: http://localhost:8025
|
|
# A distribution list that exists only inside the runner: one alias, three
|
|
# member mailboxes. Lets the fan-out case run with no external mail provider,
|
|
# no secrets and no egress.
|
|
DL_ADDRESS: ci-dl@ci.openobserve.test
|
|
DL_MEMBERS: ci-dl-m1@ci.openobserve.test,ci-dl-m2@ci.openobserve.test,ci-dl-m3@ci.openobserve.test
|
|
ZO_FORMAT_STREAM_NAME_TO_LOWERCASE: false
|
|
ZO_CREATE_ORG_THROUGH_INGESTION: true
|
|
ZO_UTF8_VIEW_ENABLED: false
|
|
ZO_ENABLE_CROSS_LINKING: true
|
|
# Timechart (visualize) tab is hidden by default; enable it so visualize E2E tests can reach it.
|
|
ZO_TIMECHART_ENABLED: true
|
|
# Allow alert destinations to target localhost so the test validation stream
|
|
# (http://localhost:5080/api/.../..._json) is not blocked by the SSRF guard.
|
|
ZO_SSRF_ALLOW_LOOPBACK: true
|
|
|
|
jobs:
|
|
check_changes:
|
|
# Runs on PRs carrying the 'ready-for-ci' label and on this repo's nightly schedule; the merge queue runs unit tests only (see docs/ci).
|
|
if: ${{ github.event_name != 'merge_group' && (github.event_name != 'schedule' || github.repository == 'openobserve/openobserve') && (github.event_name != 'pull_request' || (contains(github.event.pull_request.labels.*.name, 'ready-for-ci') && github.event.pull_request.draft == false)) }}
|
|
timeout-minutes: 5
|
|
name: check_changes
|
|
runs-on: ubuntu-latest
|
|
permissions:
|
|
contents: read
|
|
actions: read # for `gh run list` (last nightly's commit and result)
|
|
outputs:
|
|
has_changes: ${{ steps.check.outputs.has_changes }}
|
|
steps:
|
|
# Debounce: a push inside this window cancels the run (concurrency) before anything paid starts.
|
|
- if: github.event_name == 'pull_request'
|
|
run: sleep 60
|
|
- uses: actions/checkout@v5
|
|
with:
|
|
fetch-depth: 0
|
|
# Blobless: the diff needs commits and trees only, never file contents.
|
|
filter: blob:none
|
|
|
|
# One shared classification for every suite (see .github/actions/classify-changes); a change to
|
|
# this workflow file re-tests itself on top of it.
|
|
- uses: ./.github/actions/classify-changes
|
|
id: classify
|
|
# A nightly on the main commit the last nightly already passed on would only repeat its result.
|
|
- id: main_changed
|
|
if: github.event_name == 'schedule'
|
|
env:
|
|
GH_TOKEN: ${{ github.token }}
|
|
GH_REPO: ${{ github.repository }}
|
|
run: |
|
|
# On any gh failure LAST stays empty and the nightly runs (fail-safe: run rather than skip).
|
|
LAST=$(gh run list --workflow "$GITHUB_WORKFLOW" --event schedule --limit 50 \
|
|
--json databaseId,headSha,status,conclusion \
|
|
--jq "[.[] | select(.databaseId < $GITHUB_RUN_ID and .status == \"completed\")] | .[0] // empty | .headSha + \" \" + .conclusion" \
|
|
2>/dev/null || echo "")
|
|
read -r LAST_SHA LAST_CONCLUSION <<< "$LAST" || true
|
|
echo "Current main SHA: $GITHUB_SHA"
|
|
echo "Last nightly: ${LAST_SHA:-<none>} (${LAST_CONCLUSION:-<none>})"
|
|
# A red main must stay red, so only a passed nightly on the same commit lets this one skip.
|
|
if [ -n "$LAST_SHA" ] && [ "$LAST_SHA" = "$GITHUB_SHA" ] && [ "$LAST_CONCLUSION" = "success" ]; then
|
|
echo "changed=false" >> "$GITHUB_OUTPUT"
|
|
echo "::notice::main unchanged since the last nightly passed ($GITHUB_SHA); skipping."
|
|
else
|
|
echo "changed=true" >> "$GITHUB_OUTPUT"
|
|
fi
|
|
- id: check
|
|
env:
|
|
EVENT_NAME: ${{ github.event_name }}
|
|
MAIN_CHANGED: ${{ steps.main_changed.outputs.changed }}
|
|
HAS_CHANGES: ${{ steps.classify.outputs.run_playwright }}
|
|
CHANGED_WORKFLOWS: ${{ steps.classify.outputs.changed_workflows }}
|
|
run: |
|
|
if [ "$EVENT_NAME" = "schedule" ]; then
|
|
echo "has_changes=$([ "$MAIN_CHANGED" = "false" ] && echo false || echo true)" >> "$GITHUB_OUTPUT"
|
|
elif [ "$HAS_CHANGES" = "true" ] || echo " $CHANGED_WORKFLOWS " | grep -q ' playwright.yml '; then
|
|
echo "has_changes=true" >> "$GITHUB_OUTPUT"
|
|
else
|
|
echo "has_changes=false" >> "$GITHUB_OUTPUT"
|
|
fi
|
|
|
|
resolve_binary:
|
|
timeout-minutes: 5
|
|
name: resolve_binary
|
|
needs: [check_changes]
|
|
if: needs.check_changes.outputs.has_changes == 'true'
|
|
runs-on: ubuntu-latest
|
|
permissions:
|
|
actions: read
|
|
contents: read
|
|
outputs:
|
|
reuse: ${{ steps.resolve.outputs.reuse }}
|
|
run_id: ${{ steps.resolve.outputs.run_id }}
|
|
binary_artifact_name: ${{ steps.resolve.outputs.binary_artifact_name }}
|
|
frontend_artifact_name: ${{ steps.resolve.outputs.frontend_artifact_name }}
|
|
steps:
|
|
- name: Find exact-SHA Playwright runtime
|
|
id: resolve
|
|
env:
|
|
GH_TOKEN: ${{ github.token }}
|
|
SOURCE_HEAD_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
|
|
ARTIFACT_SHA: ${{ github.sha }}
|
|
CURRENT_RUN_ID: ${{ github.run_id }}
|
|
run: |
|
|
set -euo pipefail
|
|
binary_artifact="release-ci-binary-${ARTIFACT_SHA}"
|
|
frontend_artifact="frontend-dist-${ARTIFACT_SHA}"
|
|
selected_run_id="$CURRENT_RUN_ID"
|
|
reuse=false
|
|
|
|
# Scope candidates to this workflow and source head. ARTIFACT_SHA is
|
|
# the exact tested merge/commit SHA, so a base-branch update cannot
|
|
# reuse a binary built from an older synthetic merge commit.
|
|
run_ids="$(
|
|
gh api --method GET \
|
|
"repos/${GITHUB_REPOSITORY}/actions/workflows/playwright.yml/runs?head_sha=${SOURCE_HEAD_SHA}&per_page=100" \
|
|
--jq '.workflow_runs | sort_by(.created_at) | reverse | .[].id' \
|
|
2>/dev/null || true
|
|
)"
|
|
while IFS= read -r run_id; do
|
|
[ -n "$run_id" ] || continue
|
|
artifact_names="$(
|
|
gh api --method GET \
|
|
"repos/${GITHUB_REPOSITORY}/actions/runs/${run_id}/artifacts?per_page=100" \
|
|
--jq '.artifacts[] | select(.expired == false) | .name' \
|
|
2>/dev/null || true
|
|
)"
|
|
if grep -Fxq "$binary_artifact" <<< "$artifact_names" &&
|
|
grep -Fxq "$frontend_artifact" <<< "$artifact_names"; then
|
|
selected_run_id="$run_id"
|
|
reuse=true
|
|
break
|
|
fi
|
|
done <<< "$run_ids"
|
|
|
|
{
|
|
echo "reuse=$reuse"
|
|
echo "run_id=$selected_run_id"
|
|
echo "binary_artifact_name=$binary_artifact"
|
|
echo "frontend_artifact_name=$frontend_artifact"
|
|
} >> "$GITHUB_OUTPUT"
|
|
if [ "$reuse" = "true" ]; then
|
|
echo "Reusing exact-SHA Playwright runtime from run $selected_run_id"
|
|
else
|
|
echo "No reusable exact-SHA Playwright runtime found; build_binary will run"
|
|
fi
|
|
|
|
build_binary:
|
|
timeout-minutes: 45
|
|
name: build_binary
|
|
needs: [check_changes, resolve_binary]
|
|
if: >-
|
|
always() &&
|
|
needs.check_changes.result == 'success' &&
|
|
needs.check_changes.outputs.has_changes == 'true' &&
|
|
needs.resolve_binary.result == 'success' &&
|
|
needs.resolve_binary.outputs.reuse != 'true'
|
|
runs-on:
|
|
labels: ubicloud-standard-16
|
|
permissions:
|
|
contents: read
|
|
|
|
steps:
|
|
- name: Remove unused tools
|
|
run: |
|
|
sudo rm -rf /usr/share/dotnet
|
|
sudo rm -rf /opt/ghc
|
|
sudo rm -rf "/usr/local/share/boost"
|
|
sudo rm -rf "$AGENT_TOOLSDIRECTORY"
|
|
|
|
- name: Clone the current repo
|
|
uses: actions/checkout@v5
|
|
|
|
- uses: ./.github/actions/github-auth
|
|
- name: Setup Rust Toolchain
|
|
uses: dtolnay/rust-toolchain@683b588c487a69d01f6c6ddf40f036ff682cc2c8 # master
|
|
with:
|
|
toolchain: nightly-2026-10-24
|
|
targets: x86_64-unknown-linux-gnu
|
|
- uses: Swatinem/rust-cache@d1527c8f67ef9af357290475d4af4b0fab8f12e0 # v2.9.3
|
|
with:
|
|
cache-on-failure: true
|
|
shared-key: playwright-release-ci
|
|
# Save the shared cache on merge_group runs (the main-ward build now that
|
|
# push:main is gone); keep the main-ref check for any non-PR fallback.
|
|
save-if: ${{ github.event_name == 'merge_group' || github.ref == 'refs/heads/main' }}
|
|
|
|
- name: Install Protoc
|
|
run: |
|
|
bash ./.github/protoc.sh
|
|
|
|
- uses: actions/setup-node@v7
|
|
with:
|
|
node-version: 24
|
|
cache: npm
|
|
cache-dependency-path: web/package-lock.json
|
|
|
|
- name: Install frontend dependencies
|
|
run: cd web && npm ci
|
|
|
|
- name: Build binary and frontend
|
|
env:
|
|
NODE_OPTIONS: "--max-old-space-size=8192"
|
|
GIT_HASH_OVERRIDE: ${{ github.sha }}
|
|
run: |
|
|
mkdir -p /tmp/dist
|
|
bash -c "cd web && npm run build && cp -r ./dist/. /tmp/dist/" &
|
|
NPM_PID=$!
|
|
sed -i 's|folder = "../../web/dist/"|folder = "/tmp/dist/"|' src/web/src/lib.rs && grep -q 'folder = "/tmp/dist/"' src/web/src/lib.rs
|
|
cargo build --profile release-ci --target x86_64-unknown-linux-gnu
|
|
wait $NPM_PID
|
|
test -f /tmp/dist/index.html
|
|
|
|
- name: Upload binary artifact
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: release-ci-binary-${{ github.sha }}
|
|
path: target/x86_64-unknown-linux-gnu/release-ci/openobserve
|
|
retention-days: 1
|
|
|
|
- name: Upload frontend artifact
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: frontend-dist-${{ github.sha }}
|
|
path: /tmp/dist
|
|
retention-days: 1
|
|
if-no-files-found: error
|
|
|
|
# Build the shard matrix from tests/ui-testing/ci-matrix/ci_matrix.json (single source of truth,
|
|
# shared with ENT via build-ci-matrix.js). Runs under the same gate as build_binary so
|
|
# ui_integration_tests always has a matrix whenever it is allowed to run.
|
|
generate_matrix:
|
|
timeout-minutes: 5
|
|
name: generate_matrix
|
|
needs: [check_changes]
|
|
if: needs.check_changes.outputs.has_changes == 'true'
|
|
runs-on: ubuntu-latest
|
|
permissions:
|
|
contents: read
|
|
outputs:
|
|
matrix: ${{ steps.gen.outputs.matrix }}
|
|
# Kept from the earlier attempt on "Re-run failed jobs"; equal to run_attempt on "Re-run all jobs".
|
|
attempt: ${{ github.run_attempt }}
|
|
smoke_active: ${{ steps.gen.outputs.smoke_active }}
|
|
shards_selected: ${{ steps.gen.outputs.shards_selected }}
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
with:
|
|
# Full history only where the selection diff needs it; merge_group never diffs.
|
|
fetch-depth: ${{ github.event_name == 'pull_request' && '0' || '1' }}
|
|
filter: blob:none
|
|
- name: Selection contract tests
|
|
run: node --test .github/scripts/build-ci-matrix.test.js
|
|
- id: gen
|
|
env:
|
|
# Smoke selection runs on PR pushes; the merge queue no longer runs Playwright, so the
|
|
# full matrix runs once per PR when auto-merge is enabled ("Merge when ready").
|
|
# Escape hatches: the 'e2e-full' PR label forces the full matrix on every push; the
|
|
# PW_SMOKE_DISABLED repo variable ('true') turns smoke off everywhere without a code change.
|
|
IS_PR: ${{ github.event_name == 'pull_request' }}
|
|
HAS_FULL_LABEL: ${{ contains(github.event.pull_request.labels.*.name, 'e2e-full') || github.event.pull_request.auto_merge != null }}
|
|
SMOKE_DISABLED: ${{ vars.PW_SMOKE_DISABLED }}
|
|
BASE_REF: ${{ github.base_ref }}
|
|
run: |
|
|
SMOKE_DISABLED=$(echo "$SMOKE_DISABLED" | tr '[:upper:]' '[:lower:]')
|
|
SELECT_ARGS=""
|
|
if [ "$IS_PR" == "true" ] && [ "$HAS_FULL_LABEL" != "true" ] && [ "$SMOKE_DISABLED" != "true" ]; then
|
|
# Three-dot: diff from the merge base, so commits that landed on the base branch
|
|
# after this PR's merge commit was cut never leak into the selection.
|
|
# --no-renames lists a moved file under BOTH paths, so the shard owning the old
|
|
# location still runs; quotepath off keeps non-ASCII names matchable.
|
|
if git -c core.quotepath=false diff --no-renames --name-only "origin/${BASE_REF:-main}...${{ github.sha }}" > /tmp/pw-changed-files.txt; then
|
|
SELECT_ARGS="--select-for-changes /tmp/pw-changed-files.txt"
|
|
else
|
|
echo "::warning::changed-file diff failed — running the full matrix"
|
|
fi
|
|
fi
|
|
# || true: under bash -e a die() would abort before the log below ever prints.
|
|
MATRIX=$(node .github/scripts/build-ci-matrix.js tests/ui-testing/ci-matrix/ci_matrix.json $SELECT_ARGS 2> /tmp/pw-select-log.txt) || true
|
|
cat /tmp/pw-select-log.txt >&2
|
|
if [ -z "$MATRIX" ]; then echo "::error::matrix generation failed (see build-ci-matrix log above)"; exit 1; fi
|
|
echo "matrix=$MATRIX" >> "$GITHUB_OUTPUT"
|
|
COUNT=$(echo "$MATRIX" | node -e "const m=JSON.parse(require('fs').readFileSync(0,'utf8')); console.log('generated',m.include.length,'shards:', m.include.map(s=>s.testfolder).join(', ')); require('fs').appendFileSync(process.env.GITHUB_OUTPUT, 'shards_selected='+m.include.length+'\n')")
|
|
echo "smoke_active=$([ -n "$SELECT_ARGS" ] && echo true || echo false)" >> "$GITHUB_OUTPUT"
|
|
# Durable record of what ran and why, visible from the PR checks page.
|
|
{
|
|
echo "### Playwright shard selection"
|
|
echo "Mode: $([ -n "$SELECT_ARGS" ] && echo "smoke (changed-path selection)" || echo "full matrix")"
|
|
echo '```'
|
|
echo "$COUNT"
|
|
cat /tmp/pw-select-log.txt
|
|
echo '```'
|
|
} >> "$GITHUB_STEP_SUMMARY"
|
|
|
|
# One sticky comment per PR (found by the hidden marker) so the triage result updates
|
|
# in place on every push instead of spamming a fresh comment.
|
|
- name: Upsert shard-selection comment on PR
|
|
if: ${{ github.event_name == 'pull_request' }}
|
|
continue-on-error: true
|
|
env:
|
|
GH_TOKEN: ${{ secrets.ORG_ADMIN_TOKEN }} # PAT: comment as a human, not github-actions[bot]
|
|
REPO: ${{ github.repository }}
|
|
PR_NUMBER: ${{ github.event.pull_request.number }}
|
|
HEAD_SHA: ${{ github.event.pull_request.head.sha }}
|
|
MATRIX: ${{ steps.gen.outputs.matrix }}
|
|
SMOKE_ACTIVE: ${{ steps.gen.outputs.smoke_active }}
|
|
ATTEMPT: ${{ github.run_attempt }}
|
|
run: |
|
|
MARKER="<!-- playwright-shard-triage -->"
|
|
BODY=$(MARKER="$MARKER" node .github/scripts/pw-triage-comment.js /tmp/pw-select-log.txt)
|
|
CID=$(gh api "repos/$REPO/issues/$PR_NUMBER/comments" --paginate \
|
|
-q ".[] | select(.body | contains(\"$MARKER\")) | .id" | head -1)
|
|
if [ -n "$CID" ]; then
|
|
gh api -X PATCH "repos/$REPO/issues/comments/$CID" -f body="$BODY" >/dev/null && echo "Updated triage comment ($CID)."
|
|
else
|
|
gh api -X POST "repos/$REPO/issues/$PR_NUMBER/comments" -f body="$BODY" >/dev/null && echo "Posted triage comment."
|
|
fi
|
|
|
|
# A same-commit run queued behind a passed one would only repeat it; own job so build_binary never waits on generate_matrix.
|
|
prior_pass:
|
|
timeout-minutes: 5
|
|
name: prior_pass
|
|
needs: [generate_matrix]
|
|
# A manual re-run asks for the tests to run again, so only a first attempt may reuse a pass.
|
|
if: github.event_name == 'pull_request' && github.run_attempt == 1 && needs.generate_matrix.result == 'success'
|
|
runs-on: ubuntu-latest
|
|
permissions:
|
|
actions: read
|
|
outputs:
|
|
already_passed: ${{ steps.find.outputs.already_passed }}
|
|
run_url: ${{ steps.find.outputs.run_url }}
|
|
steps:
|
|
- name: Find a passed run of this exact commit
|
|
id: find
|
|
env:
|
|
GH_TOKEN: ${{ github.token }}
|
|
SOURCE_HEAD_SHA: ${{ github.event.pull_request.head.sha }}
|
|
ARTIFACT_SHA: ${{ github.sha }}
|
|
CURRENT_RUN_ID: ${{ github.run_id }}
|
|
MATRIX: ${{ needs.generate_matrix.outputs.matrix }}
|
|
run: |
|
|
set -uo pipefail
|
|
already_passed=false
|
|
run_url=""
|
|
tmp="$(mktemp -d)"
|
|
shards="$(jq -r '.include[].testfolder' <<< "$MATRIX" 2>/dev/null | sort -u || true)"
|
|
# Only completed, successful runs count: a failure is never copied and an in-flight run proves nothing.
|
|
candidates="$(
|
|
gh api --method GET \
|
|
"repos/${GITHUB_REPOSITORY}/actions/workflows/playwright.yml/runs?head_sha=${SOURCE_HEAD_SHA}&event=pull_request&status=success&per_page=100" \
|
|
--jq "[.workflow_runs[] | select(.id != ${CURRENT_RUN_ID} and .status == \"completed\" and .conclusion == \"success\")] | sort_by(.created_at) | reverse | .[] | \"\(.id) \(.run_attempt) \(.html_url)\"" \
|
|
2>/dev/null || true
|
|
)"
|
|
while read -r run_id attempt url; do
|
|
[ -n "$shards" ] && [ -n "${run_id:-}" ] || continue
|
|
passed="$(
|
|
gh api --method GET --paginate \
|
|
"repos/${GITHUB_REPOSITORY}/actions/runs/${run_id}/jobs?filter=latest&per_page=100" \
|
|
--jq '.jobs[] | select(.conclusion == "success" and (.name | startswith("e2e / "))) | .name | ltrimstr("e2e / ")' \
|
|
2>/dev/null | sort -u || true
|
|
)"
|
|
if [ -n "$(comm -23 <(printf '%s\n' "$shards") <(printf '%s\n' "$passed"))" ]; then
|
|
echo "Run $run_id did not pass every selected shard"; continue
|
|
fi
|
|
# The head SHA alone is not the tested commit: the PR merge commit moves with the base branch.
|
|
advice_id="$(
|
|
gh api --method GET \
|
|
"repos/${GITHUB_REPOSITORY}/actions/runs/${run_id}/artifacts?name=rerun-advice-attempt-${attempt}" \
|
|
--jq '[.artifacts[] | select(.expired == false)][0].id // empty' \
|
|
2>/dev/null || true
|
|
)"
|
|
[ -n "$advice_id" ] || { echo "Run $run_id has no rerun advice to prove its tested commit"; continue; }
|
|
tested_sha="$(
|
|
gh api "repos/${GITHUB_REPOSITORY}/actions/artifacts/${advice_id}/zip" > "$tmp/advice.zip" 2>/dev/null &&
|
|
unzip -p "$tmp/advice.zip" rerun-advice.json 2>/dev/null | jq -r '.tested_oss_sha // empty' 2>/dev/null || true
|
|
)"
|
|
if [ "$tested_sha" != "$ARTIFACT_SHA" ]; then
|
|
echo "Run $run_id tested ${tested_sha:-an unknown commit}, not $ARTIFACT_SHA"; continue
|
|
fi
|
|
already_passed=true
|
|
run_url="$url"
|
|
break
|
|
done <<< "$candidates"
|
|
{
|
|
echo "already_passed=$already_passed"
|
|
echo "run_url=$run_url"
|
|
} >> "$GITHUB_OUTPUT"
|
|
if [ "$already_passed" = "true" ]; then
|
|
echo "::notice::Every selected shard already passed on $ARTIFACT_SHA in $run_url; not re-running them."
|
|
else
|
|
echo "No earlier passed run covers this commit and shard selection; shards will run"
|
|
fi
|
|
|
|
ui_integration_tests:
|
|
timeout-minutes: 45
|
|
name: e2e / ${{ matrix.testfolder }}
|
|
needs: [resolve_binary, build_binary, generate_matrix, prior_pass]
|
|
if: >-
|
|
always() &&
|
|
!cancelled() &&
|
|
needs.resolve_binary.result == 'success' &&
|
|
needs.generate_matrix.result == 'success' &&
|
|
needs.prior_pass.outputs.already_passed != 'true' &&
|
|
(needs.resolve_binary.outputs.reuse == 'true' || needs.build_binary.result == 'success')
|
|
runs-on:
|
|
labels: eks-openobserve-standard-8
|
|
permissions:
|
|
contents: read
|
|
actions: read
|
|
# container:
|
|
# image: mcr.microsoft.com/playwright:v1.50.0-jammy
|
|
# options: --user root
|
|
strategy:
|
|
fail-fast: false
|
|
# Matrix is generated from tests/ui-testing/ci-matrix/ci_matrix.json by the
|
|
# generate_matrix job (the single source of truth shared with ENT). To add
|
|
# or move a spec, edit that JSON file — do NOT hand-edit a matrix here.
|
|
matrix: ${{ fromJSON(needs.generate_matrix.outputs.matrix) }}
|
|
steps:
|
|
- name: Kill background apt processes
|
|
run: |
|
|
sudo systemctl stop apt-daily.service apt-daily-upgrade.service unattended-upgrades.service 2>/dev/null || true
|
|
sudo systemctl stop apt-daily.timer apt-daily-upgrade.timer 2>/dev/null || true
|
|
sudo systemctl disable apt-daily.service apt-daily-upgrade.service unattended-upgrades.service 2>/dev/null || true
|
|
sudo flock --wait 30 /var/lib/dpkg/lock-frontend true 2>/dev/null || true
|
|
|
|
- name: Clone the current repo
|
|
uses: actions/checkout@v5
|
|
|
|
- name: Download binary artifact
|
|
uses: actions/download-artifact@v5
|
|
with:
|
|
name: ${{ needs.resolve_binary.outputs.binary_artifact_name }}
|
|
path: release-ci-binary
|
|
github-token: ${{ github.token }}
|
|
repository: ${{ github.repository }}
|
|
run-id: ${{ needs.resolve_binary.outputs.run_id }}
|
|
|
|
- name: Download frontend artifact
|
|
uses: actions/download-artifact@v5
|
|
with:
|
|
name: ${{ needs.resolve_binary.outputs.frontend_artifact_name }}
|
|
path: /tmp/dist
|
|
github-token: ${{ github.token }}
|
|
repository: ${{ github.repository }}
|
|
run-id: ${{ needs.resolve_binary.outputs.run_id }}
|
|
|
|
- name: Start mail sink (Mailpit)
|
|
# A local SMTP server the tests can also READ. Runs as a static binary
|
|
# rather than a service container: these shards run on a self-hosted EKS
|
|
# runner where Docker-in-Docker is not guaranteed, and the app itself is
|
|
# already launched the same way. ~2s, and it makes ZO_SMTP_* real.
|
|
#
|
|
# Gated to the one shard that needs it — every other shard would
|
|
# otherwise pay a GitHub-releases download for nothing.
|
|
if: matrix.testfolder == 'Alerts'
|
|
env:
|
|
# Pinned, not "latest": an upstream release must never change what this
|
|
# shard does without a commit here.
|
|
MAILPIT_VERSION: v1.21.8
|
|
run: |
|
|
set +e
|
|
curl -fsSL -o mailpit.tar.gz \
|
|
"https://github.com/axllent/mailpit/releases/download/${MAILPIT_VERSION}/mailpit-linux-amd64.tar.gz" \
|
|
&& tar -xzf mailpit.tar.gz mailpit && chmod +x ./mailpit
|
|
if [ ! -x ./mailpit ]; then
|
|
echo "::warning::Mailpit ${MAILPIT_VERSION} could not be fetched; email delivery assertions will skip"
|
|
exit 0
|
|
fi
|
|
# Loopback only — this runner is shared, and delivered alert bodies
|
|
# should not be readable from the rest of its network.
|
|
./mailpit --smtp 127.0.0.1:1025 --listen 127.0.0.1:8025 > mailpit.log 2>&1 &
|
|
for i in $(seq 1 30); do
|
|
curl -sf http://localhost:8025/api/v1/messages > /dev/null 2>&1 && { echo "Mailpit ready after ${i}s"; exit 0; }
|
|
sleep 1
|
|
done
|
|
echo "::warning::Mailpit did not start; email delivery assertions will skip"
|
|
|
|
- name: Start OpenObserve
|
|
# Per-shard override of the workflow-level default. ZO_QUICK_MODE_ENABLED is
|
|
# read once at server start and is global to the process, so a shard is the
|
|
# smallest unit that can change it. Set via "quick_mode_enabled" in
|
|
# tests/ui-testing/ci-matrix/ci_matrix.json; shards that omit it get "false".
|
|
env:
|
|
ZO_QUICK_MODE_ENABLED: ${{ matrix.quick_mode_enabled || 'false' }}
|
|
# Same per-shard rule: /api/{org}/synthetics* routes register only when true at start ("synthetics_enabled").
|
|
ZO_SYNTHETICS_ENABLED: ${{ matrix.synthetics_enabled || 'false' }}
|
|
# Same per-shard rule: how far back (hours) ingestion accepts a record.
|
|
# The workflow default of 5 DROPS older rows while still answering 200,
|
|
# so a suite that seeds history (SLO measures a rolling 7-day window)
|
|
# would fail as "No data" with nothing in the artifacts to explain it.
|
|
# Set via "ingest_allowed_upto" in ci-matrix/ci_matrix.json; shards that
|
|
# omit it keep the workflow default.
|
|
ZO_INGEST_ALLOWED_UPTO: ${{ matrix.ingest_allowed_upto || env.ZO_INGEST_ALLOWED_UPTO }}
|
|
# Same per-shard rule: how much history one SLO backfill chunk covers.
|
|
# The production default (86400 = 1 day) makes a 7-day-window test SLO
|
|
# backfill in 7 sequential chunks under ZO_SCHEDULER_SLO_BACKFILL_CONCURRENCY=1
|
|
# (SLO-Measurement's shard note above) — no test here asserts on
|
|
# per-chunk timing or partial coverage, only on the final measured
|
|
# state, so a shard-only larger chunk is free wall-clock, not a
|
|
# behavior change. Set via "slo_backfill_chunk_secs" in
|
|
# ci-matrix/ci_matrix.json; shards that omit it keep the server default.
|
|
ZO_SLO_BACKFILL_CHUNK_SECS: ${{ matrix.slo_backfill_chunk_secs || env.ZO_SLO_BACKFILL_CHUNK_SECS }}
|
|
run: chmod +x ./release-ci-binary/openobserve && ./release-ci-binary/openobserve > o2.log 2>&1 &
|
|
|
|
- name: Wait for start
|
|
run: |
|
|
for i in $(seq 1 120); do
|
|
if curl -sf http://localhost:5080/web/login > /dev/null 2>&1; then
|
|
echo "Server ready after ${i}s"
|
|
exit 0
|
|
fi
|
|
sleep 1
|
|
done
|
|
echo "Server failed to start in 120s"
|
|
exit 1
|
|
|
|
- name: Ensure we are getting a reply from the server
|
|
run: curl http://localhost:5080/web/login
|
|
|
|
- uses: actions/setup-node@v7
|
|
with:
|
|
node-version: 24
|
|
cache: npm
|
|
cache-dependency-path: tests/ui-testing/package-lock.json
|
|
|
|
- name: Start distribution-list relay
|
|
# Mailpit is a sink: it accepts any address but stores ONE message, so it
|
|
# can never show what a real distribution list does. This tiny relay sits
|
|
# in front of it and expands the alias into per-member deliveries, the
|
|
# same way alerts-content-templates.spec.js stands up an HTTP receiver to
|
|
# capture what a webhook destination really sent.
|
|
#
|
|
# Gated to the one shard that needs it, matching the Mailpit step above.
|
|
if: matrix.testfolder == 'Alerts'
|
|
run: |
|
|
if ! command -v node > /dev/null 2>&1; then
|
|
echo "::warning::node not on PATH; DL fan-out case will skip"
|
|
exit 0
|
|
fi
|
|
node tests/ui-testing/scripts/dl-relay.js > dl-relay.log 2>&1 &
|
|
for i in $(seq 1 20); do
|
|
(echo > /dev/tcp/127.0.0.1/1026) >/dev/null 2>&1 && { echo "DL relay ready after ${i}s"; exit 0; }
|
|
sleep 1
|
|
done
|
|
echo "::warning::DL relay did not start; the fan-out case will skip"
|
|
|
|
- name: Cache Playwright browser
|
|
id: playwright-cache
|
|
uses: actions/cache@v6
|
|
with:
|
|
path: ~/.cache/ms-playwright
|
|
key: playwright-chromium-${{ runner.os }}-${{ hashFiles('tests/ui-testing/package-lock.json') }}
|
|
|
|
# Expressions have no arithmetic, so the previous attempt number is computed here.
|
|
- name: Resolve previous attempt
|
|
id: prev
|
|
run: echo "attempt=$(( ${{ github.run_attempt }} - 1 ))" >> "$GITHUB_OUTPUT"
|
|
|
|
- name: Download previous attempt's failed-test list
|
|
if: ${{ github.run_attempt > 1 }}
|
|
continue-on-error: true
|
|
uses: actions/download-artifact@v5
|
|
with:
|
|
name: last-run-${{ matrix.testfolder }}-attempt-${{ steps.prev.outputs.attempt }}
|
|
path: tests/ui-testing/last-run-prev
|
|
|
|
- name: Install dependencies and run ui-tests
|
|
run: |
|
|
touch .env
|
|
echo "ZO_ROOT_USER_EMAIL=${ZO_ROOT_USER_EMAIL}" >> .env
|
|
echo "ZO_ROOT_USER_PASSWORD=${ZO_ROOT_USER_PASSWORD}" >> .env
|
|
echo "ZO_BASE_URL=${ZO_BASE_URL}" >> .env
|
|
echo "WS_ZO_BASE_URL=${WS_ZO_BASE_URL}" >> .env
|
|
echo "ZO_BASE_URL_SC=${ZO_BASE_URL_SC}" >> .env
|
|
echo "ZO_BASE_URL_SC_UI=${ZO_BASE_URL_SC_UI}" >> .env
|
|
echo "INGESTION_URL=${INGESTION_URL}" >> .env
|
|
echo "ORGNAME=${ORGNAME}" >> .env
|
|
echo "ZO_SMTP_ENABLED=${ZO_SMTP_ENABLED}" >> .env
|
|
echo "MAILPIT_URL=${MAILPIT_URL}" >> .env
|
|
echo "DL_ADDRESS=${DL_ADDRESS}" >> .env
|
|
echo "DL_MEMBERS=${DL_MEMBERS}" >> .env
|
|
mv .env tests/ui-testing
|
|
cd tests/ui-testing && npm ci
|
|
|
|
# Install apt system dependencies (always needed).
|
|
# azure.archive.ubuntu.com intermittently times out fetching font/lib
|
|
# packages, which makes `install-deps` exit 100 and fails the whole job.
|
|
# Give apt its own per-fetch retries and wrap the call in a backoff loop.
|
|
echo 'Acquire::Retries "5";' | sudo tee /etc/apt/apt.conf.d/99retries >/dev/null
|
|
for attempt in 1 2 3; do
|
|
if npx playwright install-deps chromium; then
|
|
echo "install-deps succeeded on attempt ${attempt}"
|
|
break
|
|
fi
|
|
if [ "${attempt}" = 3 ]; then
|
|
echo "::error::playwright install-deps chromium failed after 3 attempts"
|
|
exit 1
|
|
fi
|
|
echo "install-deps attempt ${attempt} failed; retrying in $((attempt * 20))s"
|
|
sleep $((attempt * 20))
|
|
done
|
|
|
|
# Get revision from playwright's own metadata
|
|
CHROMIUM_REV=$(node -pe "
|
|
require('./node_modules/playwright-core/browsers.json')
|
|
.browsers.find(b => b.name === 'chromium').revision
|
|
")
|
|
|
|
install_browser() {
|
|
local NAME="$1"
|
|
local BINARY="$2"
|
|
local ZIP_PATH="$3"
|
|
local DIR="$HOME/.cache/ms-playwright/${NAME}-${CHROMIUM_REV}"
|
|
if [ -f "${DIR}/INSTALLATION_COMPLETE" ]; then
|
|
echo "${NAME} already installed (cache hit)"; return 0
|
|
fi
|
|
rm -rf "${DIR}"
|
|
echo "Cache miss — downloading ${NAME} r${CHROMIUM_REV} via curl + system unzip"
|
|
curl -fL --progress-bar \
|
|
"https://cdn.playwright.dev/dbazure/download/playwright/builds/${ZIP_PATH}" \
|
|
-o "/tmp/${NAME}.zip"
|
|
mkdir -p "${DIR}"
|
|
unzip -q "/tmp/${NAME}.zip" -d "${DIR}"
|
|
rm -f "/tmp/${NAME}.zip"
|
|
find "${DIR}" -name "${BINARY}" -type f -exec chmod 755 {} \;
|
|
find "${DIR}" -name 'chrome_sandbox' -type f -exec chmod 4755 {} \; 2>/dev/null || true
|
|
touch "${DIR}/INSTALLATION_COMPLETE"
|
|
echo "${NAME} installed: $(ls ${DIR}/)"
|
|
}
|
|
|
|
install_browser "chromium" "chrome" \
|
|
"chromium/${CHROMIUM_REV}/chromium-linux.zip"
|
|
install_browser "chromium_headless_shell" "headless_shell" \
|
|
"chromium/${CHROMIUM_REV}/chromium-headless-shell-linux.zip"
|
|
|
|
# Get list of files to run using join function
|
|
FILE_LIST="${{ join(matrix.run_files, ' ') }}"
|
|
echo "DEBUG: FILE_LIST = $FILE_LIST"
|
|
echo "DEBUG: matrix.testfolder = ${{ matrix.testfolder }}"
|
|
|
|
if [ -n "$FILE_LIST" ]; then
|
|
# actual_folder (the real directory under playwright-tests/) comes straight
|
|
# from the matrix now — the logical->directory mapping lives in ci-matrix/ci_matrix.json.
|
|
ACTUAL_FOLDER="${{ matrix.actual_folder }}"
|
|
echo "DEBUG: ACTUAL_FOLDER = $ACTUAL_FOLDER"
|
|
|
|
# Build file paths
|
|
FILE_PATHS=""
|
|
for file in $FILE_LIST; do
|
|
FILE_PATH="./playwright-tests/$ACTUAL_FOLDER/$file"
|
|
echo "DEBUG: Will run file: $FILE_PATH"
|
|
FILE_PATHS="$FILE_PATHS $FILE_PATH"
|
|
done
|
|
|
|
# Optional per-shard worker pin. Empty keeps playwright.config.js's
|
|
# default; a shard whose specs contend for a shared server-side
|
|
# resource sets "workers": 1 in ci_matrix.json, because fullyParallel
|
|
# races separate spec FILES and mode:'serial' only orders tests
|
|
# within one file.
|
|
WORKERS_ARG=""
|
|
if [ -n "${{ matrix.workers }}" ]; then
|
|
WORKERS_ARG="--workers=${{ matrix.workers }}"
|
|
echo "DEBUG: pinning workers to ${{ matrix.workers }}"
|
|
fi
|
|
|
|
# Re-run whole files that had a failure (a serial group needs its earlier tests).
|
|
PREV=last-run-prev/.last-run.json
|
|
if [ "$GITHUB_RUN_ATTEMPT" -gt 1 ] && [ "${{ needs.generate_matrix.outputs.attempt }}" = "$GITHUB_RUN_ATTEMPT" ]; then
|
|
echo "Re-run all jobs: running the full shard"
|
|
elif [ "$GITHUB_RUN_ATTEMPT" -gt 1 ] && [ -f "$PREV" ] \
|
|
&& node -e "const r=require('./$PREV'); process.exit(r.status === 'failed' && (r.failedTests||[]).length ? 0 : 1)"; then
|
|
mkdir -p test-results && cp "$PREV" test-results/.last-run.json
|
|
LIST_JSON=$(mktemp)
|
|
read -ra SHARD_FILES <<< "$FILE_PATHS"
|
|
FAILED_FILES=""
|
|
if PLAYWRIGHT_JSON_OUTPUT_NAME="$LIST_JSON" npx playwright test "${SHARD_FILES[@]}" --last-failed --list --reporter=json > /dev/null; then
|
|
FAILED_FILES=$(node -e "const r=JSON.parse(require('fs').readFileSync(process.argv[1], 'utf8')); console.log([...new Set((r.suites||[]).map(s => s.file).filter(Boolean))].sort().map(f => './playwright-tests/' + f).join(' '))" "$LIST_JSON" || true)
|
|
fi
|
|
# An attempt that dies before writing its own list must not upload the previous attempt's as its own.
|
|
rm -f test-results/.last-run.json
|
|
if [ -n "$FAILED_FILES" ]; then
|
|
echo "Re-run: files with failed tests from attempt ${{ steps.prev.outputs.attempt }}: $FAILED_FILES"
|
|
FILE_PATHS="$FAILED_FILES"
|
|
else
|
|
echo "::warning::Could not list the failed tests from attempt ${{ steps.prev.outputs.attempt }} — running the full shard"
|
|
fi
|
|
elif [ "$GITHUB_RUN_ATTEMPT" -gt 1 ]; then
|
|
echo "::warning::No complete failed-test list from attempt ${{ steps.prev.outputs.attempt }} — running the full shard"
|
|
fi
|
|
|
|
echo "DEBUG: Final command: npx playwright test $FILE_PATHS $WORKERS_ARG"
|
|
npx playwright test $FILE_PATHS $WORKERS_ARG
|
|
else
|
|
echo "No files specified to run for ${{ matrix.testfolder }} folder"
|
|
fi
|
|
|
|
- name: Upload blob report to GitHub Actions Artifacts
|
|
if: ${{ !cancelled() }}
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: blob-report-${{ matrix.testfolder }}-attempt-${{ github.run_attempt }}
|
|
path: tests/ui-testing/blob-report
|
|
retention-days: 1
|
|
|
|
# job-status lets merge_reports tell a shard that died before writing results from one that passed.
|
|
- name: Record shard result
|
|
if: ${{ !cancelled() }}
|
|
continue-on-error: true
|
|
env:
|
|
JOB_STATUS: ${{ job.status }}
|
|
run: |
|
|
cd tests/ui-testing
|
|
rm -rf shard-record && mkdir shard-record
|
|
printf '%s\n' "$JOB_STATUS" > shard-record/job-status
|
|
if [ -f test-results/.last-run.json ]; then cp test-results/.last-run.json shard-record/; fi
|
|
|
|
- name: Upload failed-test list for re-runs
|
|
if: ${{ !cancelled() }}
|
|
continue-on-error: true
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: last-run-${{ matrix.testfolder }}-attempt-${{ github.run_attempt }}
|
|
path: tests/ui-testing/shard-record
|
|
include-hidden-files: true
|
|
if-no-files-found: ignore
|
|
retention-days: 1
|
|
|
|
- name: Show OpenObserve log tail
|
|
if: failure()
|
|
run: tail -n 200 o2.log || true
|
|
|
|
- name: Upload OpenObserve log
|
|
if: failure()
|
|
continue-on-error: true
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: o2-log-${{ matrix.testfolder }}-attempt-${{ github.run_attempt }}
|
|
path: o2.log
|
|
retention-days: 1
|
|
if-no-files-found: ignore
|
|
|
|
merge_reports:
|
|
timeout-minutes: 20
|
|
name: Merge Reports
|
|
needs: [generate_matrix, ui_integration_tests]
|
|
if: ${{ !cancelled() && needs.ui_integration_tests.result != 'skipped' }}
|
|
# eks runner: `playwright merge-reports` silently stops mid-merge (exits 0, ~3 of N blobs, no
|
|
# output) on the ubicloud and github-hosted runners — but works on the eks runners (same as ENT).
|
|
runs-on:
|
|
labels: eks-openobserve-standard-4
|
|
permissions:
|
|
contents: read
|
|
actions: read # the no-blobs investigation reads run/job state from the Actions API
|
|
outputs:
|
|
# Compact Playwright stats (expected/unexpected/flaky/skipped) from the merged report.json,
|
|
# consumed by report_to_openobserve to record per-test counts in ci_test_runs.
|
|
stats: ${{ steps.report_stats.outputs.stats }}
|
|
# Flaky + failed tests for the ci_test_results stream, consumed by
|
|
# report_to_openobserve. JSON array: [{module,title,file,status}] where
|
|
# status is "flaky"|"failed" (see .github/scripts/extract-flaky.js). "[]" if none.
|
|
flaky: ${{ steps.report_flaky.outputs.flaky }}
|
|
|
|
steps:
|
|
- name: Clone the current repo
|
|
uses: actions/checkout@v5
|
|
|
|
- name: Setup Node.js
|
|
uses: actions/setup-node@v7
|
|
with:
|
|
node-version: 24
|
|
cache: npm
|
|
cache-dependency-path: tests/ui-testing/package-lock.json
|
|
|
|
- name: Install project dependencies
|
|
run: cd tests/ui-testing && npm ci
|
|
|
|
- name: Download blob reports from GitHub Actions Artifacts
|
|
uses: actions/download-artifact@v4
|
|
continue-on-error: true
|
|
id: download_blobs
|
|
with:
|
|
path: tests/ui-testing/all-blob-reports
|
|
pattern: blob-report-*-attempt-${{ github.run_attempt }}
|
|
merge-multiple: true
|
|
|
|
- name: Check if any artifacts were downloaded
|
|
run: |
|
|
if [ -d "tests/ui-testing/all-blob-reports" ] && [ "$(ls -A tests/ui-testing/all-blob-reports)" ]; then
|
|
echo "✅ Artifacts downloaded successfully"
|
|
ls -la tests/ui-testing/all-blob-reports
|
|
else
|
|
echo "⚠️ No artifacts found or download failed completely"
|
|
# Don't fail here - let the Merge Reports step handle this
|
|
fi
|
|
|
|
- name: Merge Reports
|
|
env:
|
|
GH_TOKEN: ${{ github.token }}
|
|
run: |
|
|
cd tests/ui-testing
|
|
|
|
# Check if directory exists and contains reports
|
|
if [ ! -d "all-blob-reports" ] || [ -z "$(ls -A all-blob-reports 2>/dev/null)" ]; then
|
|
echo "::warning::No blob reports found to merge."
|
|
|
|
# Investigate: Check if all test jobs succeeded (meaning they skipped intentionally)
|
|
# vs test jobs failed (meaning blob reporter might have failed)
|
|
echo "::group::Investigating why no blob reports exist"
|
|
|
|
JOBS_RESPONSE=$(curl -s -H "Authorization: token $GH_TOKEN" \
|
|
-H "Accept: application/vnd.github.v3+json" \
|
|
"https://api.github.com/repos/${{ github.repository }}/actions/runs/${{ github.run_id }}/attempts/${{ github.run_attempt }}/jobs")
|
|
|
|
# Count test matrix jobs by conclusion
|
|
SUCCEEDED=$(echo "$JOBS_RESPONSE" | jq -r '[.jobs[] | select(.name | startswith("e2e /")) | select(.conclusion == "success")] | length')
|
|
FAILED=$(echo "$JOBS_RESPONSE" | jq -r '[.jobs[] | select(.name | startswith("e2e /")) | select(.conclusion == "failure")] | length')
|
|
TOTAL=$(echo "$JOBS_RESPONSE" | jq -r '[.jobs[] | select(.name | startswith("e2e /"))] | length')
|
|
|
|
echo "::notice::Test job results: $SUCCEEDED succeeded, $FAILED failed, $TOTAL total"
|
|
|
|
if [ "$SUCCEEDED" -gt 0 ] && [ "$FAILED" -eq 0 ]; then
|
|
echo "::notice::✅ All test jobs succeeded - they likely skipped because tests already passed"
|
|
echo "::notice::This is VALID behavior for optimized reruns"
|
|
echo "::notice::Creating empty report structure for downstream jobs"
|
|
mkdir -p playwright-results/html-report
|
|
# Add stats object so downstream report consumers get a well-formed report.
|
|
# startTime is an ISO 8601 string (not a Unix timestamp number).
|
|
START_TIME=$(date -u +%Y-%m-%dT%H:%M:%S.000Z)
|
|
echo '{"config":{},"suites":[],"errors":[],"stats":{"startTime":"'"$START_TIME"'","duration":0,"expected":0,"unexpected":0,"flaky":0,"skipped":0}}' > playwright-results/report.json
|
|
echo "<html><body><h1>No tests executed - all shards skipped</h1><p>All test shards exited early because tests had already passed in a previous attempt.</p></body></html>" > playwright-results/html-report/index.html
|
|
echo "::endgroup::"
|
|
exit 0
|
|
else
|
|
echo "::error::❌ Some/all test jobs failed but no blob reports were found"
|
|
echo "::error::This indicates a problem with test execution or blob reporter"
|
|
echo "::error::Failed jobs: $FAILED, Succeeded jobs: $SUCCEEDED"
|
|
echo "::endgroup::"
|
|
exit 1
|
|
fi
|
|
fi
|
|
|
|
# Force the html + json reporters explicitly with their output paths instead of
|
|
# relying on `unset CI` to flip playwright.config.js's `process.env.CI ? blob : html/json`
|
|
# ternary. The runner exports CI=true and `unset CI` does not reliably reach the
|
|
# merge subprocess, so the config would intermittently re-select the `blob` reporter —
|
|
# which writes a merged blob-report/ and NOTHING to playwright-results/, making the
|
|
# downstream `ls playwright-results/html-report/` fail with a cryptic exit 2.
|
|
mkdir -p playwright-results
|
|
# These env vars override playwright.config.js's reporter settings:
|
|
# PLAYWRIGHT_HTML_OUTPUT_DIR - html reporter output directory
|
|
# PLAYWRIGHT_HTML_OPEN=never - never auto-open the report (required for CI)
|
|
# PLAYWRIGHT_JSON_OUTPUT_NAME - json reporter output file path
|
|
PLAYWRIGHT_HTML_OUTPUT_DIR=playwright-results/html-report \
|
|
PLAYWRIGHT_HTML_OPEN=never \
|
|
PLAYWRIGHT_JSON_OUTPUT_NAME=playwright-results/report.json \
|
|
npx playwright merge-reports --reporter html,json ./all-blob-reports
|
|
|
|
echo "Contents of playwright-results:"
|
|
ls -la playwright-results/ || true
|
|
if [ ! -f playwright-results/html-report/index.html ] || [ ! -f playwright-results/report.json ]; then
|
|
echo "::error::merge-reports did not produce the expected html-report/index.html and report.json"
|
|
echo "::error::all-blob-reports contents:"
|
|
ls -la all-blob-reports/ || true
|
|
exit 1
|
|
fi
|
|
# Diagnostics only — never fail the step on these (the guard above already
|
|
# confirmed the outputs exist).
|
|
echo "Contents of html-report:"
|
|
ls -la playwright-results/html-report/ || true
|
|
echo "Checking report.json:"
|
|
ls -lh playwright-results/report.json || true
|
|
|
|
- name: Extract report stats for OpenObserve
|
|
id: report_stats
|
|
if: always()
|
|
run: |
|
|
cd tests/ui-testing
|
|
if [ -f playwright-results/report.json ]; then
|
|
STATS=$(jq -c '.stats // {}' playwright-results/report.json 2>/dev/null || echo '{}')
|
|
else
|
|
STATS='{}'
|
|
fi
|
|
echo "stats=$STATS" >> "$GITHUB_OUTPUT"
|
|
echo "report stats: $STATS"
|
|
|
|
- name: Extract flaky tests for OpenObserve
|
|
id: report_flaky
|
|
if: always()
|
|
run: |
|
|
# Flaky + failed tests from the merged report, as one JSON array, for the
|
|
# ci_test_results stream. Logic + schema live in extract-flaky.js (it
|
|
# handles a missing/malformed report by printing [] non-fatally, and is
|
|
# commented to explain status="flaky" vs "failed" and the sp.title choice).
|
|
FLAKY=$(node .github/scripts/extract-flaky.js tests/ui-testing/playwright-results/report.json)
|
|
echo "flaky=$FLAKY" >> "$GITHUB_OUTPUT"
|
|
# Log a count only — avoid dumping test file paths/titles into public CI logs.
|
|
echo "flaky/failed test count: $(printf '%s' "$FLAKY" | jq 'length' 2>/dev/null || echo '?')"
|
|
|
|
# Read by playwright-auto-rerun.yml; failed_count -1 means "no advice", so a missing or broken report still writes the file.
|
|
- name: Write rerun advice
|
|
if: always()
|
|
env:
|
|
TESTED_OSS_SHA: ${{ github.sha }}
|
|
run: |
|
|
mkdir -p rerun-advice
|
|
# Report file paths are relative to playwright.config.js testDir.
|
|
jq -ce --arg sha "$TESTED_OSS_SHA" '
|
|
if (.suites | type) != "array" then error("not a Playwright report") else . end
|
|
| [.suites[] | recurse(.suites[]?) | . as $s | (.specs // [])[] | . as $sp
|
|
| (.tests // [])[] | select(.status == "unexpected") | ($sp.file // $s.file // "")]
|
|
| {tested_oss_sha: $sha, failed_count: length,
|
|
failed_files: (map(select(. != "") | "tests/ui-testing/playwright-tests/" + .) | unique)}
|
|
' tests/ui-testing/playwright-results/report.json > rerun-advice/rerun-advice.json 2>/dev/null \
|
|
|| jq -nc --arg sha "$TESTED_OSS_SHA" '{tested_oss_sha: $sha, failed_count: -1, failed_files: []}' > rerun-advice/rerun-advice.json
|
|
jq -c '{failed_count, failed_file_count: (.failed_files | length)}' rerun-advice/rerun-advice.json
|
|
|
|
# runner.temp is emptied per job, so a reused runner cannot supply an earlier run's records.
|
|
- name: Download failed-test lists
|
|
id: download_last_runs
|
|
if: ${{ always() && github.run_attempt == 1 && needs.ui_integration_tests.result != 'success' }}
|
|
continue-on-error: true
|
|
uses: actions/download-artifact@v4
|
|
with:
|
|
path: ${{ runner.temp }}/last-runs
|
|
pattern: last-run-*-attempt-${{ github.run_attempt }}
|
|
|
|
# A shard that died before writing results adds nothing to the report, so the advice would miss its failure.
|
|
- name: Check shard results match the merged report
|
|
if: always()
|
|
env:
|
|
SHARDS_RESULT: ${{ needs.ui_integration_tests.result }}
|
|
LAST_RUNS_OUTCOME: ${{ steps.download_last_runs.outcome }}
|
|
MATRIX: ${{ needs.generate_matrix.outputs.matrix }}
|
|
run: |
|
|
ADVICE=rerun-advice/rerun-advice.json
|
|
COMPUTED=$(cat "$ADVICE" 2>/dev/null || true)
|
|
COUNT=$(jq '.failed_count' <<< "$COMPUTED" 2>/dev/null) && [ "$COUNT" -ge 0 ] 2>/dev/null || exit 0
|
|
# The bot reads advice only for a failed run.
|
|
[ "$SHARDS_RESULT" != "success" ] || exit 0
|
|
# The file says "no advice" until every check below passes, so any early exit keeps the bot's old behaviour.
|
|
jq -c '.failed_count = -1 | .failed_files = []' <<< "$COMPUTED" > "$ADVICE"
|
|
if [ "$GITHUB_RUN_ATTEMPT" != "1" ]; then
|
|
# The bot reads only attempt 1's advice, and a failed-jobs rerun leaves no record for shards that passed earlier.
|
|
echo "Rerun advice is only written for the first attempt"; exit 0
|
|
fi
|
|
if [ "$LAST_RUNS_OUTCOME" != "success" ]; then
|
|
echo "::warning::Could not download the shard results; no rerun advice"; exit 0
|
|
fi
|
|
if ! SHARDS=$(jq -r '.include | if type == "array" and length > 0 and all(.testfolder | type == "string") then .[].testfolder else error("not a shard matrix") end' <<< "$MATRIX" 2>/dev/null) \
|
|
|| [ -z "$SHARDS" ]; then
|
|
echo "::warning::Could not read the shard list from the matrix; no rerun advice"; exit 0
|
|
fi
|
|
LISTED=0
|
|
FAILED_SHARDS=""
|
|
while IFS= read -r SHARD; do
|
|
RECORD="$RUNNER_TEMP/last-runs/last-run-$SHARD-attempt-$GITHUB_RUN_ATTEMPT"
|
|
if ! STATUS=$(cat "$RECORD/job-status" 2>/dev/null); then
|
|
echo "::warning::Shard $SHARD left no result record; no rerun advice"; exit 0
|
|
fi
|
|
if [ "$STATUS" != "success" ]; then
|
|
jq -e '.status == "failed" and (.failedTests | type == "array" and length > 0)' "$RECORD/.last-run.json" > /dev/null 2>&1 \
|
|
|| { echo "::warning::Shard $SHARD failed without a complete failed-test list; no rerun advice"; exit 0; }
|
|
FAILED_SHARDS="$FAILED_SHARDS $SHARD"
|
|
fi
|
|
if [ -f "$RECORD/.last-run.json" ]; then
|
|
N=$(jq -e '.failedTests | if type == "array" then length else error("no failedTests array") end' "$RECORD/.last-run.json" 2>/dev/null) \
|
|
|| { echo "::warning::Shard $SHARD left an unreadable failed-test list; no rerun advice"; exit 0; }
|
|
LISTED=$((LISTED + N))
|
|
fi
|
|
done <<< "$SHARDS"
|
|
# Catches a failed shard whose blob report never reached the merged report.
|
|
if [ "$LISTED" -ne "$COUNT" ]; then
|
|
echo "::warning::Failed-test lists hold $LISTED tests but the merged report has $COUNT; no rerun advice"; exit 0
|
|
fi
|
|
printf '%s\n' "$COMPUTED" > "$ADVICE"
|
|
echo "rerun advice kept; failed shards checked:$FAILED_SHARDS"
|
|
|
|
- name: Upload rerun advice
|
|
if: always()
|
|
continue-on-error: true
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: rerun-advice-attempt-${{ github.run_attempt }}
|
|
path: rerun-advice/rerun-advice.json
|
|
retention-days: 7
|
|
|
|
- name: Upload merged report
|
|
if: always()
|
|
uses: actions/upload-artifact@v4
|
|
with:
|
|
name: merged-report-${{ github.run_attempt }}
|
|
path: tests/ui-testing/playwright-results
|
|
retention-days: 1
|
|
if-no-files-found: warn
|
|
|
|
# OpenObserve metrics — its own job so playwright_summary never depends on it. Emits one
|
|
# retry-aware run-summary doc (ci_test_runs, suite=ui) + per-test flaky/failed rows
|
|
# (ci_test_results). The step is continue-on-error and the poster scripts never fail.
|
|
report_openobserve:
|
|
timeout-minutes: 10
|
|
name: Report to OpenObserve
|
|
needs: [generate_matrix, ui_integration_tests, merge_reports]
|
|
if: ${{ !cancelled() && needs.ui_integration_tests.result != 'skipped' }}
|
|
runs-on:
|
|
labels: eks-openobserve-standard-4
|
|
permissions:
|
|
contents: read
|
|
actions: read # reads run/job timings from the Actions API
|
|
steps:
|
|
- name: Clone the current repo
|
|
uses: actions/checkout@v5
|
|
|
|
- name: Report run metrics to OpenObserve
|
|
continue-on-error: true
|
|
env:
|
|
GH_TOKEN: ${{ github.token }}
|
|
O2_REPORTING_INGEST_BASE: ${{ secrets.O2_REPORTING_INGEST_BASE }}
|
|
O2_REPORTING_AUTH: ${{ secrets.O2_REPORTING_AUTH }}
|
|
O2_REPORTING_INSECURE: ${{ vars.O2_REPORTING_INSECURE }}
|
|
# Track lightweight readiness separately from a successful fresh
|
|
# build; skipped reuse-path builds are reported as null.
|
|
RUNTIME_READY_JOB_NAME: resolve_binary
|
|
TESTS: ${{ needs.ui_integration_tests.result }}
|
|
MERGE_STATS: ${{ needs.merge_reports.outputs.stats }}
|
|
MERGE_FLAKY: ${{ needs.merge_reports.outputs.flaky }}
|
|
# Branch + PR for per-test rows so the dashboard needs no cross-stream JOIN.
|
|
# head_ref is set on pull_request (the source branch); ref_name covers push +
|
|
# merge_group (gh-readonly-queue/main/pr-<n>-<sha>). PR empty on merge_group/push.
|
|
REPORT_BRANCH: ${{ github.head_ref || github.ref_name }}
|
|
REPORT_PR: ${{ github.event.pull_request.number }}
|
|
SMOKE_ACTIVE: ${{ needs.generate_matrix.outputs.smoke_active }}
|
|
SHARDS_SELECTED: ${{ needs.generate_matrix.outputs.shards_selected }}
|
|
run: |
|
|
# This job runs only when tests actually ran (merge `if:` gates on !cancelled &&
|
|
# ui != skipped), so TESTS is success|failure here.
|
|
if [ "$TESTS" = "failure" ]; then CONCLUSION=failure; else CONCLUSION=success; fi
|
|
# Per-test counts from the merged report.json (tests_total includes skipped).
|
|
if [ -n "${MERGE_STATS:-}" ] && [ "$MERGE_STATS" != "null" ] && [ "$MERGE_STATS" != "{}" ]; then
|
|
EXTRA_JSON=$(printf '%s' "$MERGE_STATS" | jq -c '{tests_total:((.expected//0)+(.unexpected//0)+(.flaky//0)+(.skipped//0)), tests_passed:(.expected//0), tests_failed:(.unexpected//0), tests_flaky:(.flaky//0), tests_skipped:(.skipped//0)}' 2>/dev/null || echo "")
|
|
fi
|
|
# smoke_active/shards_selected let dashboards separate smoke rows from full runs,
|
|
# since shards_total now varies by selection instead of always being the full set.
|
|
SMOKE_JSON=$(jq -nc --arg a "${SMOKE_ACTIVE:-false}" --arg n "${SHARDS_SELECTED:-}" \
|
|
'{smoke_active:($a=="true"), shards_selected:(($n|tonumber?)//null)}')
|
|
if [ -n "${EXTRA_JSON:-}" ]; then
|
|
EXTRA_JSON=$(jq -nc --argjson a "$EXTRA_JSON" --argjson b "$SMOKE_JSON" '$a * $b')
|
|
else
|
|
EXTRA_JSON="$SMOKE_JSON"
|
|
fi
|
|
export EXTRA_JSON
|
|
export CONCLUSION STREAM=ci_test_runs SUITE=ui
|
|
bash .github/scripts/o2-run-report.sh
|
|
# One row per flaky-or-failed test -> ci_test_results (status flaky|failed; suite/
|
|
# workflow hardcoded UI-only). o2-report.sh takes the stream as arg 1. Non-fatal.
|
|
FLAKY="${MERGE_FLAKY:-[]}"
|
|
if printf '%s' "$FLAKY" | jq -e 'type=="array" and length>0' >/dev/null 2>&1; then
|
|
FLAKY_FILE=$(mktemp "${TMPDIR:-/tmp}/flaky.XXXXXX")
|
|
trap 'rm -f "$FLAKY_FILE"' EXIT
|
|
printf '%s' "$FLAKY" | jq -c --arg run_id "$GITHUB_RUN_ID" --arg run_attempt "$GITHUB_RUN_ATTEMPT" --arg repo "$GITHUB_REPOSITORY" \
|
|
--arg branch "${REPORT_BRANCH:-}" --arg pr "${REPORT_PR:-}" \
|
|
'[.[] | {run_id:$run_id, run_attempt:($run_attempt|tonumber), repo:$repo, suite:"ui", workflow:"test", ingest_source:"live", status:.status, module:.module, test_title:.title, test_file:.file, branch:(if $branch=="" then null else $branch end), pr_number:(if $pr=="" then null else ($pr|tonumber) end)}]' \
|
|
> "$FLAKY_FILE" 2>/dev/null || echo '[]' > "$FLAKY_FILE"
|
|
if [ "$(jq 'length' "$FLAKY_FILE" 2>/dev/null || echo 0)" -gt 0 ]; then
|
|
bash .github/scripts/o2-report.sh ci_test_results "$FLAKY_FILE"
|
|
fi
|
|
fi
|
|
|
|
playwright_summary:
|
|
timeout-minutes: 5
|
|
runs-on:
|
|
labels: eks-openobserve-standard-4
|
|
permissions:
|
|
checks: write
|
|
pull-requests: read
|
|
needs:
|
|
[
|
|
check_changes,
|
|
resolve_binary,
|
|
build_binary,
|
|
generate_matrix,
|
|
prior_pass,
|
|
ui_integration_tests,
|
|
]
|
|
if: always() && github.event_name != 'merge_group' && (github.event_name != 'schedule' || github.repository == 'openobserve/openobserve') && (github.event_name != 'pull_request' || (contains(github.event.pull_request.labels.*.name, 'ready-for-ci') && github.event.pull_request.draft == false))
|
|
steps:
|
|
- name: Check test results
|
|
id: verdict
|
|
env:
|
|
ALREADY_PASSED: ${{ needs.prior_pass.result == 'success' && needs.prior_pass.outputs.already_passed == 'true' }}
|
|
PRIOR_RUN_URL: ${{ needs.prior_pass.outputs.run_url }}
|
|
run: |
|
|
# 1. check_changes must have succeeded — if it failed, downstream jobs skipped
|
|
# and we must not report a false success.
|
|
if [ "${{ needs.check_changes.result }}" != "success" ]; then
|
|
echo "check_changes job did not succeed: ${{ needs.check_changes.result }}"
|
|
exit 1
|
|
fi
|
|
|
|
# 2. Nothing Playwright-relevant changed → nothing ran → allow.
|
|
if [ "${{ needs.check_changes.outputs.has_changes }}" != "true" ]; then
|
|
echo "✅ No Playwright-relevant changes — suite not required."
|
|
exit 0
|
|
fi
|
|
|
|
# prior_pass matched the tested commit and every selected shard, so the earlier verdict stands.
|
|
if [ "$ALREADY_PASSED" = "true" ]; then
|
|
if [ "${{ needs.build_binary.result }}" = "failure" ]; then
|
|
echo "build_binary failed on a commit that already passed: ${{ needs.build_binary.result }}"
|
|
exit 1
|
|
fi
|
|
echo "✅ Every selected shard already passed on this commit in $PRIOR_RUN_URL"
|
|
exit 0
|
|
fi
|
|
|
|
# 3. Suite was expected to run. The resolver must succeed, and either
|
|
# an exact-SHA pair was reused or this run built a fresh pair.
|
|
if [ "${{ needs.resolve_binary.result }}" != "success" ]; then
|
|
echo "resolve_binary job did not succeed: ${{ needs.resolve_binary.result }}"
|
|
exit 1
|
|
fi
|
|
if [ "${{ needs.resolve_binary.outputs.reuse }}" != "true" ] && \
|
|
[ "${{ needs.build_binary.result }}" != "success" ]; then
|
|
echo "No reusable artifact and build_binary did not succeed: ${{ needs.build_binary.result }}"
|
|
exit 1
|
|
fi
|
|
|
|
# 4. Runtime readiness alone is not enough; all selected shards must pass.
|
|
if [ "${{ needs.ui_integration_tests.result }}" == "success" ]; then
|
|
echo "All Playwright tests completed successfully"
|
|
exit 0
|
|
else
|
|
echo "Playwright tests failed:"
|
|
echo " resolve_binary: ${{ needs.resolve_binary.result }} (reuse=${{ needs.resolve_binary.outputs.reuse }})"
|
|
echo " build_binary: ${{ needs.build_binary.result }}"
|
|
echo " ui_integration_tests: ${{ needs.ui_integration_tests.result }}"
|
|
exit 1
|
|
fi
|
|
|
|
- name: Complete the auto-merge placeholder
|
|
if: ${{ !cancelled() && github.event_name == 'pull_request' && github.event.pull_request.head.repo.fork == false && github.actor != 'dependabot[bot]' && needs.check_changes.result == 'success' }}
|
|
# github-script, not gh: the eks runners are not known to ship the gh CLI.
|
|
uses: actions/github-script@v9
|
|
env:
|
|
VERDICT: ${{ steps.verdict.outcome }}
|
|
SMOKE: ${{ needs.check_changes.outputs.has_changes == 'true' && needs.generate_matrix.outputs.smoke_active == 'true' }}
|
|
HAS_CHANGES: ${{ needs.check_changes.outputs.has_changes }}
|
|
with:
|
|
script: |
|
|
if (process.env.SMOKE === 'true') {
|
|
// A smoke verdict may stand in only once auto-merge is off, else a full run dropped from the queue strands the placeholder.
|
|
const { data: pr } = await github.rest.pulls.get({ ...context.repo, pull_number: context.payload.pull_request.number });
|
|
if (pr.auto_merge) { core.info('Smoke run and auto-merge is on: leaving the placeholder to the full-matrix run.'); return; }
|
|
}
|
|
const conclusion = process.env.VERDICT === 'success' ? 'success' : 'failure';
|
|
const scope = process.env.HAS_CHANGES !== 'true' ? 'No Playwright changes' : process.env.SMOKE === 'true' ? 'Smoke' : 'Full matrix';
|
|
const runUrl = `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`;
|
|
const runs = await github.paginate(github.rest.checks.listForRef, {
|
|
...context.repo, ref: context.payload.pull_request.head.sha, check_name: 'playwright_summary', filter: 'all', per_page: 100,
|
|
});
|
|
const placeholders = runs.filter((r) => r.external_id === 'playwright-automerge-placeholder');
|
|
if (!placeholders.length) core.info('No auto-merge placeholder on this SHA.');
|
|
for (const r of placeholders) {
|
|
await github.rest.checks.update({
|
|
...context.repo, check_run_id: r.id, status: 'completed', conclusion, details_url: runUrl,
|
|
output: { title: `${scope}: ${conclusion}`, summary: `Completed by the ${scope} run ${runUrl}` },
|
|
});
|
|
core.info(`placeholder ${r.id} -> ${conclusion}`);
|
|
}
|