Skip to content
E2E Nightly

E2E Nightly #83

Sign in to view logs

E2E Nightly

E2E Nightly #83

Workflow file for this run

.github/workflows/e2e-nightly.yaml at 27f1ed2
name: E2E Nightly
on:
schedule:
# Run at 3:00 AM UTC every day
- cron: "0 3 * * *"
workflow_dispatch:
inputs:
k8s-version:
description: Kubernetes version to test
required: true
type: choice
default: all
options:
- all
- v1.32
- v1.33
- v1.34
- v1.35
suite:
description: Test suite to run
required: true
type: choice
default: all
options:
- all
- chainsaw
- go-e2e
permissions:
contents: read
concurrency:
group: e2e-nightly-${{ github.ref }}-${{ github.event_name }}-${{ inputs.k8s-version || 'all' }}-${{ inputs.suite || 'all' }}
cancel-in-progress: true
env:
K3D_VERSION: "v5.8.3"
CERT_MANAGER_VERSION: "v1.17.2"
PROMETHEUS_IMAGE: "quay.io/prometheus/prometheus:v3.4.1"
PROMETHEUS_CHART_VERSION: "27.22.0"
STRESS_NG_IMAGE: "ghcr.io/alexei-led/stress-ng:0.20.01"
CPU_BURN_IMAGE: "docker.io/library/busybox:1.37"
jobs:
prepare-matrix:
name: Select nightly matrix
runs-on: ${{ vars.RUNNER || 'ubuntu-latest' }}
timeout-minutes: 5
outputs:
matrix: ${{ steps.set-matrix.outputs.matrix }}
suite: ${{ steps.set-matrix.outputs.suite }}
steps:
- name: Select version and suite
id: set-matrix
shell: bash -Eeuo pipefail {0}
run: |
selected_version="${{ inputs.k8s-version || 'all' }}"
selected_suite="${{ inputs.suite || 'all' }}"
case "$selected_version" in
v1.32)
matrix='{"include":[{"k8s-version":"v1.32","k3s-image":"v1.32.13-k3s1"}]}'
;;
v1.33)
matrix='{"include":[{"k8s-version":"v1.33","k3s-image":"v1.33.11-k3s1"}]}'
;;
v1.34)
matrix='{"include":[{"k8s-version":"v1.34","k3s-image":"v1.34.7-k3s1"}]}'
;;
v1.35)
matrix='{"include":[{"k8s-version":"v1.35","k3s-image":"v1.35.4-k3s1"}]}'
;;
all)
matrix='{"include":[{"k8s-version":"v1.32","k3s-image":"v1.32.13-k3s1"},{"k8s-version":"v1.33","k3s-image":"v1.33.11-k3s1"},{"k8s-version":"v1.34","k3s-image":"v1.34.7-k3s1"},{"k8s-version":"v1.35","k3s-image":"v1.35.4-k3s1"}]}'
;;
*)
echo "Unknown k8s-version input: $selected_version" >&2
exit 1
;;
esac
echo "matrix=$matrix" >> "$GITHUB_OUTPUT"
echo "suite=$selected_suite" >> "$GITHUB_OUTPUT"
test-e2e:
name: E2E (K8s ${{ matrix.k8s-version }})
if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
needs: prepare-matrix
runs-on: ${{ vars.RUNNER || 'ubuntu-latest' }}
timeout-minutes: 75
env:
SELECTED_SUITE: ${{ needs.prepare-matrix.outputs.suite }}
CLUSTER_NAME: e2e-${{ matrix.k8s-version }}-${{ github.run_id }}-${{ github.run_attempt }}
KUBECONFIG: ${{ github.workspace }}/.kubeconfig-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.k8s-version }}-${{ needs.prepare-matrix.outputs.suite }}
strategy:
max-parallel: 4
fail-fast: false
matrix: ${{ fromJSON(needs.prepare-matrix.outputs.matrix) }}
steps:
- uses: step-security/harden-runner@9af89fc71515a100421586dfdb3dc9c984fbf411 # v2.19.4
with:
egress-policy: audit
- uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
- uses: ./.github/actions/setup-clean-docker-config
- name: Verify Docker and buildx
shell: bash -Eeuo pipefail {0}
run: |
echo "Docker version: $(docker version --format '{{.Client.Version}}')"
echo "Docker config: ${DOCKER_CONFIG:-~/.docker}"
if ! docker buildx version >/dev/null 2>&1; then
echo "::error::Docker buildx is not available. Check that cli-plugins are accessible from DOCKER_CONFIG."
docker buildx version || true
ls -la "${DOCKER_CONFIG:-$HOME/.docker}/cli-plugins/" 2>/dev/null || echo "No cli-plugins directory found"
exit 1
fi
echo "buildx: $(docker buildx version)"
- uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6
with:
go-version-file: go.mod
cache: true
- name: Cache E2E container images
uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
with:
path: |
/tmp/cert-manager-controller.tar
/tmp/cert-manager-webhook.tar
/tmp/cert-manager-cainjector.tar
/tmp/cert-manager.yaml
/tmp/prometheus.tar
/tmp/stress-ng.tar
/tmp/busybox.tar
/tmp/k3s.tar
key: e2e-images-${{ runner.os }}-cm${{ env.CERT_MANAGER_VERSION }}-prom${{ env.PROMETHEUS_CHART_VERSION }}-stress0.20.01-busybox1.37-k3s${{ matrix.k3s-image }}
- uses: ./.github/actions/install-binary-tool
with:
name: k3d
version: ${{ env.K3D_VERSION }}
install-command: |
curl -s https://raw.githubusercontent.com/k3d-io/k3d/${{ env.K3D_VERSION }}/install.sh | K3D_INSTALL_DIR="$TOOL_DIR" USE_SUDO=false bash
- name: Cleanup stale k3d clusters and Docker resources
shell: bash -Eeuo pipefail -x {0}
run: |
# Delete any k3d cluster that has been running for more than 1 hour.
# Stale clusters from crashed/cancelled jobs waste Docker resources
# and cause container name conflicts during creation.
for cluster in $(k3d cluster list -o json 2>/dev/null | jq -r '.[].name // empty'); do
# Get the server container start time; skip if we cannot determine age.
started=$(docker inspect --format '{{.State.StartedAt}}' "k3d-${cluster}-server-0" 2>/dev/null) || continue
started_epoch=$(date -d "$started" +%s 2>/dev/null) || continue
age=$(( $(date +%s) - started_epoch ))
if (( age > 3600 )); then
echo "Deleting stale k3d cluster '$cluster' (age: ${age}s)"
k3d cluster delete "$cluster" || true
fi
done
docker builder prune -f || true
docker system prune -f || true
- name: Pre-pull container images
shell: bash -Eeuo pipefail -x {0}
run: |
# Pull images with docker (parallel layer downloads) and save as
# tarballs for k3d import. Skip images already restored from cache.
# Accepts an optional 3rd argument: a fallback image reference to try
# when the primary registry is unreachable (e.g. quay.io outage).
# If the fallback succeeds, the image is retagged to the primary name
# so downstream manifests (cert-manager.yaml) resolve correctly.
pull_and_save() {
local image="$1" tarball="$2" fallback="${3:-}"
[[ -f "$tarball" ]] && { echo "Cached: $tarball"; return 0; }
for attempt in 1 2 3; do
if docker pull --platform linux/amd64 "$image"; then
docker save "$image" -o "$tarball"
return 0
fi
echo "::warning::docker pull attempt $attempt for $image failed, retrying in 10s..."
sleep 10
done
if [[ -n "$fallback" ]]; then
echo "::warning::Primary registry failed for $image, trying fallback: $fallback"
for attempt in 1 2 3; do
if docker pull --platform linux/amd64 "$fallback"; then
docker tag "$fallback" "$image"
docker save "$image" -o "$tarball"
return 0
fi
echo "::warning::fallback pull attempt $attempt for $fallback failed, retrying in 10s..."
sleep 10
done
fi
echo "::error::Failed to pull $image after all attempts"
return 1
}
CM_VER="${{ env.CERT_MANAGER_VERSION }}"
OWNER="attune-io"
pull_and_save "quay.io/jetstack/cert-manager-controller:${CM_VER}" /tmp/cert-manager-controller.tar "ghcr.io/${OWNER}/mirrors/cert-manager-controller:${CM_VER}"
pull_and_save "quay.io/jetstack/cert-manager-webhook:${CM_VER}" /tmp/cert-manager-webhook.tar "ghcr.io/${OWNER}/mirrors/cert-manager-webhook:${CM_VER}"
pull_and_save "quay.io/jetstack/cert-manager-cainjector:${CM_VER}" /tmp/cert-manager-cainjector.tar "ghcr.io/${OWNER}/mirrors/cert-manager-cainjector:${CM_VER}"
PROM_TAG="${{ env.PROMETHEUS_IMAGE }}"
PROM_TAG="${PROM_TAG##*:}"
pull_and_save "${{ env.PROMETHEUS_IMAGE }}" /tmp/prometheus.tar "docker.io/prom/prometheus:${PROM_TAG}"
pull_and_save "${{ env.STRESS_NG_IMAGE }}" /tmp/stress-ng.tar
pull_and_save "${{ env.CPU_BURN_IMAGE }}" /tmp/busybox.tar
pull_and_save "rancher/k3s:${{ matrix.k3s-image }}" /tmp/k3s.tar
- name: Create k3d cluster
shell: bash -Eeuo pipefail -x {0}
run: |
rm -f "$KUBECONFIG"
# Load the pre-pulled k3s image into Docker so k3d uses it
# without pulling from Docker Hub (avoids rate-limit/timeout failures).
docker load -i /tmp/k3s.tar
# Retry cluster creation up to 3 times. k3d can fail transiently
# when Docker containers vanish mid-creation due to daemon
# contention with concurrent CI runs.
#
# K8s 1.32 feature gate: InPlacePodVerticalScaling is alpha and off
# by default. We enable it via a k3s config file mounted into the
# server container, NOT via --k3s-arg CLI flags.
#
# Why config file instead of --k3s-arg:
# k3s processes --kube-apiserver-arg flags through GetArgs() in
# pkg/util/args.go. CLI args have a timing race where GetArgs()
# may process extras before defaults are populated (issue #296).
# The config file is loaded during server bootstrap, reducing
# but not eliminating the race (issue #325 showed += still fails
# intermittently even via config file).
#
# We use = (set) not += (append) to avoid the GetArgs() timing
# race entirely. The = syntax sets the feature gate directly
# without depending on any existing value being populated first.
# This is safe for K8s 1.32: k3s's default apiserver feature
# gates at this version are all GA (locked on/off, not present
# in --feature-gates), so there are no alpha/beta defaults to
# preserve.
EXTRA_K3D_ARGS=()
if [[ "${{ matrix.k8s-version }}" == "v1.32" ]]; then
K3S_CONFIG="/tmp/k3s-v132-config.yaml"
cat > "$K3S_CONFIG" <<'K3SEOF'
kube-apiserver-arg:
- "feature-gates=InPlacePodVerticalScaling=true"
kube-controller-manager-arg:
- "feature-gates=InPlacePodVerticalScaling=true"
kube-scheduler-arg:
- "feature-gates=InPlacePodVerticalScaling=true"
kubelet-arg:
- "feature-gates=InPlacePodVerticalScaling=true"
K3SEOF
EXTRA_K3D_ARGS=(--volume "${K3S_CONFIG}:/etc/rancher/k3s/config.yaml@server:*")
fi
for attempt in 1 2 3; do
if ! k3d cluster create "$CLUSTER_NAME" \
--image rancher/k3s:${{ matrix.k3s-image }} \
--k3s-arg "--disable=traefik,servicelb@server:*" \
"${EXTRA_K3D_ARGS[@]}" \
--wait --timeout 120s \
--kubeconfig-update-default=false \
--kubeconfig-switch-context=false; then
echo "::warning::k3d cluster create attempt $attempt failed, retrying in 10s..."
k3d cluster delete "$CLUSTER_NAME" 2>/dev/null || true
sleep 10
if (( attempt == 3 )); then
echo "::error::k3d cluster creation failed after 3 attempts"
exit 1
fi
continue
fi
break
done
k3d kubeconfig merge "$CLUSTER_NAME" --output "$KUBECONFIG" --overwrite
- name: Wait for node Ready
shell: bash -Eeuo pipefail -x {0}
run: kubectl wait --for=condition=Ready nodes --all --timeout=360s
- name: Verify resize subresource (K8s 1.32)
if: matrix.k8s-version == 'v1.32'
shell: bash -Eeuo pipefail -x {0}
run: |
echo "Verifying InPlacePodVerticalScaling feature gate is active..."
# Verify the config file was loaded by checking the apiserver process.
echo "k3s config inside container:"
docker exec "k3d-${CLUSTER_NAME}-server-0" cat /etc/rancher/k3s/config.yaml || true
for i in $(seq 1 24); do
if kubectl get --raw /api/v1 | jq -e '.resources[] | select(.name == "pods/resize")' >/dev/null 2>&1; then
echo "pods/resize subresource is registered"
exit 0
fi
echo "Waiting for pods/resize subresource (attempt $i/24)..."
sleep 5
done
echo "::error::pods/resize subresource not found after 120s. The InPlacePodVerticalScaling feature gate did not take effect."
echo "Diagnostic: k3s config file:"
docker exec "k3d-${CLUSTER_NAME}-server-0" cat /etc/rancher/k3s/config.yaml || true
echo "Diagnostic: k3s process (apiserver runs embedded):"
docker exec "k3d-${CLUSTER_NAME}-server-0" sh -c "ps aux | grep k3s" || true
echo "Diagnostic: API resources containing 'resize':"
kubectl get --raw /api/v1 | jq '[.resources[] | select(.name | contains("resize"))]' || true
echo "Diagnostic: k3s logs (last 100 lines):"
docker logs "k3d-${CLUSTER_NAME}-server-0" --tail 100 2>&1 | grep -iE 'feature|gate|resize|error|fatal|panic' || true
exit 1
- name: Load cert-manager images into cluster
shell: bash -Eeuo pipefail -x {0}
run: |
k3d image import \
/tmp/cert-manager-controller.tar \
/tmp/cert-manager-webhook.tar \
/tmp/cert-manager-cainjector.tar \
-c "$CLUSTER_NAME"
- name: Download cert-manager manifest
shell: bash -Eeuo pipefail -x {0}
run: |
if [[ -f /tmp/cert-manager.yaml ]]; then
echo "Cached: /tmp/cert-manager.yaml"
else
for attempt in 1 2 3; do
if curl -fsSL -o /tmp/cert-manager.yaml \
"https://github.com/cert-manager/cert-manager/releases/download/${{ env.CERT_MANAGER_VERSION }}/cert-manager.yaml"; then
break
fi
echo "::warning::cert-manager.yaml download attempt $attempt failed, retrying in 10s..."
sleep 10
if (( attempt == 3 )); then
echo "::error::Failed to download cert-manager.yaml after 3 attempts"
exit 1
fi
done
fi
- name: Install cert-manager
shell: bash -Eeuo pipefail -x {0}
run: |
kubectl apply -f /tmp/cert-manager.yaml
kubectl wait --for=condition=Available deployment/cert-manager -n cert-manager --timeout=120s
kubectl wait --for=condition=Available deployment/cert-manager-webhook -n cert-manager --timeout=120s
kubectl wait --for=condition=Available deployment/cert-manager-cainjector -n cert-manager --timeout=120s
- name: Load Prometheus and test images into cluster
shell: bash -Eeuo pipefail -x {0}
run: k3d image import /tmp/prometheus.tar /tmp/stress-ng.tar /tmp/busybox.tar -c "$CLUSTER_NAME"
- name: Install Prometheus
shell: bash -Eeuo pipefail -x {0}
run: |
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update
# Pre-pull the chart tarball with retry to avoid transient CDN failures
# (e.g. 504 Gateway Timeout from GitHub-hosted chart assets).
for attempt in 1 2 3; do
if helm pull prometheus-community/prometheus \
--version ${{ env.PROMETHEUS_CHART_VERSION }} \
--destination /tmp; then
break
fi
echo "::warning::helm pull attempt $attempt failed, retrying in 10s..."
sleep 10
if (( attempt == 3 )); then
echo "::error::Failed to pull Prometheus chart after 3 attempts"
exit 1
fi
done
helm install prometheus /tmp/prometheus-${{ env.PROMETHEUS_CHART_VERSION }}.tgz \
--namespace monitoring --create-namespace \
--set server.image.repository=quay.io/prometheus/prometheus \
--set server.image.tag=v3.4.1 \
--set server.persistentVolume.enabled=false \
--set alertmanager.enabled=false \
--set prometheus-pushgateway.enabled=false \
--set server.global.scrape_interval=15s \
--wait --timeout 5m
- name: Wait for cAdvisor metrics in Prometheus
shell: bash -Eeuo pipefail -x {0}
run: |
echo "Waiting for Prometheus to scrape cAdvisor metrics..."
PROM_POD=$(kubectl get pods -n monitoring -l app.kubernetes.io/name=prometheus,app.kubernetes.io/component=server -o name | head -1)
for i in $(seq 1 30); do
result=$(kubectl exec -n monitoring "$PROM_POD" -- \
wget -qO- 'http://localhost:9090/api/v1/query?query=container_cpu_usage_seconds_total' 2>/dev/null || true)
if echo "$result" | grep -q '"result":\[{'; then
echo "cAdvisor metrics available after ${i}x5s"
break
fi
if [ "$i" -eq 30 ]; then
echo "WARNING: cAdvisor metrics not found after 150s, proceeding anyway"
fi
sleep 5
done
- uses: ./.github/actions/install-go-tool
with:
name: ko
version: v0.18.0
package: github.com/google/ko
- name: Build and load operator image
shell: bash -Eeuo pipefail -x {0}
run: |
# Use ko to build the image as an OCI tarball. This bypasses the
# Docker daemon entirely, avoiding containerd storage races.
VERSION=e2e COMMIT=$(git rev-parse --short HEAD) DATE=$(date -u +%Y-%m-%dT%H:%M:%SZ) \
KO_DOCKER_REPO=attune ko build ./cmd/manager/ \
--bare --tags=e2e --platform=linux/$(go env GOARCH) \
--tarball=/tmp/attune-e2e.tar --push=false
k3d image import /tmp/attune-e2e.tar -c "$CLUSTER_NAME"
- name: Install operator via Helm
shell: bash -Eeuo pipefail -x {0}
run: |
helm install attune ./charts/attune \
--namespace attune-system --create-namespace \
--set image.repository=attune \
--set image.tag=e2e \
--set image.pullPolicy=Never \
--set webhooks.enabled=true \
--set metrics.enabled=true \
--set leaderElection.enabled=false \
--set maxConcurrentReconciles=4 \
--set logging.level=1 \
--set resources.limits.memory=512Mi \
--set resources.requests.memory=256Mi \
--wait --timeout 3m
- name: Install Chainsaw
if: inputs.suite != 'go-e2e'
uses: ./.github/actions/install-binary-tool
with:
name: chainsaw
version: v0.2.15
install-command: |
ARCH=$(uname -m)
case "$ARCH" in
x86_64) ARCH=amd64 ;;
aarch64) ARCH=arm64 ;;
esac
URL="https://github.com/kyverno/chainsaw/releases/download/v0.2.15/chainsaw_linux_${ARCH}.tar.gz"
for attempt in 1 2 3; do
if curl -fsSL "$URL" -o chainsaw.tar.gz; then
tar -xf chainsaw.tar.gz -C "$TOOL_DIR" chainsaw
rm chainsaw.tar.gz
chmod +x "$TOOL_DIR/chainsaw"
exit 0
fi
echo "::warning::Chainsaw download attempt $attempt failed, retrying in 10s..."
sleep 10
done
echo "::error::Failed to download Chainsaw after 3 attempts"
exit 1
- name: Run Chainsaw suite
if: inputs.suite != 'go-e2e'
shell: bash -Eeuo pipefail -x {0}
run: |
mkdir -p test-results
chainsaw test test/e2e/ --config .chainsaw.yaml 2>&1 | tee test-results/chainsaw-${{ matrix.k8s-version }}.log
- name: Run Go E2E tests
if: inputs.suite != 'chainsaw'
shell: bash -Eeuo pipefail -x {0}
run: |
mkdir -p test-results
go test -tags=e2e ./test/e2e-go/... -race -count=1 -timeout=15m -v 2>&1 | tee test-results/go-e2e-${{ matrix.k8s-version }}.log
- name: Collect debug info on failure
if: failure()
shell: bash -Eeuo pipefail {0}
run: |
mkdir -p test-results
{
echo "=== cluster name ==="
echo "$CLUSTER_NAME"
echo "=== kubeconfig path ==="
echo "$KUBECONFIG"
echo "=== kubeconfig server ==="
kubectl config view --minify --raw -o jsonpath='{.clusters[0].cluster.server}' || true
printf '\n'
echo "=== cert-manager pods ==="
kubectl get pods -n cert-manager || true
echo "=== operator logs ==="
kubectl logs -n attune-system -l app.kubernetes.io/name=attune --tail=300 || true
echo "=== pod status ==="
kubectl get pods -A || true
echo "=== pod descriptions ==="
kubectl describe pods -A || true
echo "=== events (attune-system) ==="
kubectl get events -n attune-system --sort-by='.lastTimestamp' || true
echo "=== events (all namespaces, last 100) ==="
kubectl get events -A --sort-by='.lastTimestamp' | tail -100 || true
} | tee test-results/debug-${{ matrix.k8s-version }}.log
- name: Upload test artifacts
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: e2e-nightly-${{ matrix.k8s-version }}-${{ env.SELECTED_SUITE }}
path: test-results/
if-no-files-found: ignore
retention-days: 7
- name: Cleanup k3d cluster
if: always()
shell: bash -Eeuo pipefail {0}
run: |
k3d cluster delete "$CLUSTER_NAME" 2>/dev/null || true
rm -f "$KUBECONFIG"
fuzz:
name: Fuzz Tests
runs-on: ${{ vars.RUNNER || 'ubuntu-latest' }}
timeout-minutes: 20
if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
steps:
- uses: step-security/harden-runner@9af89fc71515a100421586dfdb3dc9c984fbf411 # v2.19.4
with:
egress-policy: audit
- uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
- uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6
with:
go-version-file: go.mod
cache: true
- name: Run fuzz tests (coverage-guided, 30s per target)
shell: bash -Eeuo pipefail -x {0}
run: |
# Time-based limits let Go's coverage-guided engine focus on
# inputs that explore new code paths. 30s per target achieves
# full branch coverage faster than 1M random iterations.
go test ./internal/recommendation/... -run='^$' -fuzz=FuzzPercentileEstimator -fuzztime=30s
go test ./internal/recommendation/... -run='^$' -fuzz=FuzzRecommendationEngine -fuzztime=30s
go test ./internal/webhook/... -run='^$' -fuzz=FuzzValidateFloatFields -fuzztime=30s
report:
name: Nightly Results
runs-on: ${{ vars.RUNNER || 'ubuntu-latest' }}
timeout-minutes: 5
permissions:
issues: write
needs: [test-e2e, fuzz]
if: always() && (github.event_name == 'schedule' || github.event_name == 'workflow_dispatch')
steps:
- uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
- name: Summary
shell: bash -Eeuo pipefail {0}
run: |
e2e_result="${{ needs.test-e2e.result }}"
fuzz_result="${{ needs.fuzz.result }}"
failed=0
if [[ "$e2e_result" != "success" && "$e2e_result" != "skipped" ]]; then
echo "::error::E2E nightly failed (result: $e2e_result)"
failed=1
fi
if [[ "$fuzz_result" != "success" && "$fuzz_result" != "skipped" ]]; then
echo "::error::Fuzz nightly failed (result: $fuzz_result)"
failed=1
fi
if (( failed )); then
exit 1
fi
echo "All nightly tests passed (E2E: $e2e_result, Fuzz: $fuzz_result)"
- name: Create issue on failure
if: failure() && github.event_name == 'schedule'
env:
GH_TOKEN: ${{ github.token }}
shell: bash -Eeuo pipefail {0}
run: |
# Only create if no open issue with this label already exists.
existing=$(gh issue list --label e2e-nightly-failure --state open --json number --jq 'length')
if [[ "$existing" -gt 0 ]]; then
echo "Open nightly failure issue already exists, skipping creation"
exit 0
fi
body=$(cat <<EOF
The [nightly run](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}) failed.
- E2E result: \`${{ needs.test-e2e.result }}\`
- Fuzz result: \`${{ needs.fuzz.result }}\`
Check the workflow run for details.
EOF
)
gh issue create \
--title "Nightly failed on $(date -u +%Y-%m-%d)" \
--label e2e-nightly-failure \
--body "$body"