From 84c880513571c3f20b043a592d94f9a3f949b1fc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adam=20=C5=9Awiderski?= <adam.swiderski@hyland.com>
Date: Thu, 3 Sep 2026 15:24:34 +0200
Subject: [PATCH] [ACS-12735] Added cluster logs for failed "deploy local acs"
 step (#5370)

* [ACS-12735] Added cluster logs for failed "deploy local acs" step

* [ACS-12735] Added changes from acs-deployment file

* [ACS-12735] added pods logs to upload artifacts

* [ACS-12735] deleted duplicated code
---
 .github/actions/deploy-local-acs/action.yml   | 65 +++++++++++++++++++
 .github/workflows/e2e-playwright-reusable.yml |  8 +--
 .github/workflows/pull-request.yml            |  8 +--
 3 files changed, 67 insertions(+), 14 deletions(-)

diff --git a/.github/actions/deploy-local-acs/action.yml b/.github/actions/deploy-local-acs/action.yml
index 793addd88..5fb24ffcc 100644
--- a/.github/actions/deploy-local-acs/action.yml
+++ b/.github/actions/deploy-local-acs/action.yml
@@ -31,6 +31,13 @@ inputs:
       and PVCs to become ready, in Helm's duration format (e.g. '5m', '10m').
     required: false
     default: '10m'
+  log_name_identifier:
+    description: >-
+      Suffix that makes the uploaded pod-logs artifact name unique. Callers in a
+      matrix must pass a value unique per leg (e.g. browser + suite) to avoid
+      artifact name collisions.
+    required: false
+    default: 'deploy-local-acs'
 
 runs:
   using: "composite"
@@ -200,6 +207,56 @@ runs:
           --values acs-deployment/test/enterprise-integration-test-values.yaml \
           --values .github/acs-deployment-values-override.yaml
 
+    - name: Dump cluster state on deploy failure
+      if: failure()
+      shell: bash
+      run: |
+        echo "::group::Helm releases"
+        helm ls --all-namespaces --all || true
+        echo "::endgroup::"
+
+        echo "::group::Helm status (acs)"
+        helm status acs --show-resources || true
+        echo "::endgroup::"
+
+        echo "::group::All resources (all namespaces)"
+        kubectl get all --all-namespaces || true
+        echo "::endgroup::"
+
+        echo "::group::Node metrics"
+        kubectl top nodes || true
+        echo "::endgroup::"
+
+        echo "::group::Pod metrics (by memory)"
+        kubectl top pod --all-namespaces --sort-by=memory || true
+        echo "::endgroup::"
+
+        echo "::group::Recent cluster events"
+        kubectl get events -A --sort-by=.lastTimestamp | tail -n 100 || true
+        echo "::endgroup::"
+
+        # Filter on the Ready condition, not phase: the usual culprit is a repository pod stuck phase=Running/Ready=False on a startup-probe timeout, which a phase filter would miss.
+        pods=$(kubectl get pods -A \
+          -o jsonpath='{range .items[*]}{.metadata.namespace}{"\t"}{.metadata.name}{"\t"}{.status.phase}{"\t"}{range .status.conditions[?(@.type=="Ready")]}{.status}{end}{"\n"}{end}' \
+          2>/dev/null || true)
+        not_ready=$(echo "$pods" | awk -F'\t' '$3 != "Succeeded" && $4 != "True" { print $1" "$2 }')
+
+        if [ -z "$not_ready" ]; then
+          echo "No not-ready pods were found."
+          exit 0
+        fi
+
+        while read -r ns pod; do
+          [ -z "$ns" ] && continue
+          echo "::group::describe ${ns}/${pod}"
+          kubectl describe pod -n "$ns" "$pod" || true
+          echo "::endgroup::"
+
+          echo "::group::logs ${ns}/${pod} (previous)"
+          kubectl logs -n "$ns" "$pod" --all-containers --previous --tail=200 2>/dev/null || echo "(no previous container logs)"
+          echo "::endgroup::"
+        done <<< "$not_ready"
+
     - name: Report deployed ACS versions
       if: always()
       shell: bash
@@ -252,3 +309,11 @@ runs:
         echo "::group::Actual container images running in the cluster"
         echo "${pod_images:-(no pods found)}"
         echo "::endgroup::"
+
+    - name: Upload pods logs
+      if: always()
+      uses: Alfresco/alfresco-build-tools/.github/actions/kubectl-keep-nslogs@d2981070adf1a68c3ddcbf0ebba193467261c9b3 # v18.26.0
+      with:
+        namespace: default
+        log_retention: 7
+        log_name_identifier: "logs-pods-${{ inputs.log_name_identifier }}"
diff --git a/.github/workflows/e2e-playwright-reusable.yml b/.github/workflows/e2e-playwright-reusable.yml
index 195908fb7..4c7483287 100644
--- a/.github/workflows/e2e-playwright-reusable.yml
+++ b/.github/workflows/e2e-playwright-reusable.yml
@@ -136,13 +136,7 @@ jobs:
           quay_username: ${{ secrets.QUAY_USERNAME }}
           quay_password: ${{ secrets.QUAY_PASSWORD }}
           acs_deployment_version: 'master'
-
-      - name: Collect and push logs as artifact
-        if: always()
-        uses: Alfresco/alfresco-build-tools/.github/actions/kubectl-keep-nslogs@d2981070adf1a68c3ddcbf0ebba193467261c9b3 # v18.26.0
-        with:
-          log_retention: 7
-          log_name_identifier: "logs-${{ matrix.browser }}-${{ matrix.e2e-suites.name }}"
+          log_name_identifier: "${{ matrix.browser }}-${{ matrix.e2e-suites.name }}"
 
       - name: Before install
         uses: ./.github/actions/before-install
diff --git a/.github/workflows/pull-request.yml b/.github/workflows/pull-request.yml
index a1891cf8f..a74116ad0 100644
--- a/.github/workflows/pull-request.yml
+++ b/.github/workflows/pull-request.yml
@@ -233,13 +233,7 @@ jobs:
           quay_username: ${{ secrets.QUAY_USERNAME }}
           quay_password: ${{ secrets.QUAY_PASSWORD }}
           acs_deployment_version: 'master'
-
-      - name: Collect and push logs as artifact
-        if: always()
-        uses: Alfresco/alfresco-build-tools/.github/actions/kubectl-keep-nslogs@d2981070adf1a68c3ddcbf0ebba193467261c9b3 # v18.26.0
-        with:
-          log_retention: 7
-          log_name_identifier: "logs-${{ matrix.browser || 'chromium' }}-${{ matrix.e2e-suites.name }}"
+          log_name_identifier: "${{ matrix.browser || 'chromium' }}-${{ matrix.e2e-suites.name }}"
 
       - name: Before install
         uses: ./.github/actions/before-install