update from bundle-kubeflow

.github/workflows/deploy-to-aks.yaml

@@ -6,6 +6,9 @@ on:
         description: 'Comma-separated list of bundle versions e.g. "1.7","1.8"'
         default: '"1.8"'
         required: true
+      k8s_version:
+        description: 'Kubernetes version to be used for the AKS cluster'
+        required: false
 
   # schedule:
   #   - cron: "17 0 * * 2"
@@ -14,17 +17,17 @@ jobs:
     runs-on: ubuntu-22.04
     strategy:
       matrix:
-        bundle_version: ${{ fromJSON(format('[{0}]', inputs.bundle_version || '"1.7","1.8","latest"')) }}
+        bundle_version: ${{ fromJSON(format('[{0}]', inputs.bundle_version || '"1.8","latest"')) }}
       fail-fast: false
     env:
       AZURE_CORE_OUTPUT: none
-      K8S_VERSION: ${{ fromJSON('{"1.8":"1.26"}')[matrix.bundle_version] }}
-      JUJU_VERSION: ${{ fromJSON('{"1.8":"3.1"}')[ matrix.bundle_version ] }}
+      K8S_VERSION: ${{ inputs.k8s_version || fromJSON('{"1.8":"1.26", "latest":"1.26"}')[matrix.bundle_version] }}
+      JUJU_VERSION: ${{ fromJSON('{"1.8":"3.1","latest":"3.1"}')[ matrix.bundle_version ] }}
     steps:
       - name: Checkout repository
         uses: actions/checkout@v2
 
-      - name: Install CLI tools tox charmcraft juju
+      - name: Install CLI tools
         run: |
           python -m pip install --upgrade pip
           pip install tox
@@ -70,29 +73,38 @@ jobs:
           tox -vve test_bundle_deployment-${{ matrix.bundle_version }} -- --model kubeflow --keep-models -vv -s
 
       # On failure, capture debugging resources
+      - name: Save debug artifacts
+        uses: canonical/kubeflow-ci/actions/dump-charm-debug-artifacts@main
+        if: always()
+
       - name: Get juju status
         run: juju status
-        if: failure()
+        if: failure() || cancelled()
 
       - name: Get juju debug logs
         run: juju debug-log --replay --no-tail
-        if: failure()
+        if: failure() || cancelled()
 
       - name: Get all kubernetes resources
         run: kubectl get all -A
-        if: failure()
+        if: failure() || cancelled()
+
+      - name: Describe all pods
+        if: failure() || cancelled()
+        run: |
+          sudo microk8s kubectl describe pods --all-namespaces
 
       - name: Get logs from pods with status = Pending
         run: kubectl -n kubeflow get pods | tail -n +2 | grep Pending | awk '{print $1}' | xargs -n1 kubectl -n kubeflow logs --all-containers=true --tail 100
-        if: failure()
+        if: failure() || cancelled()
 
       - name: Get logs from pods with status = Failed
         run: kubectl -n kubeflow get pods | tail -n +2 | grep Failed | awk '{print $1}' | xargs -n1 kubectl -n kubeflow logs --all-containers=true --tail 100
-        if: failure()
+        if: failure() || cancelled()
 
       - name: Get logs from pods with status = CrashLoopBackOff
         run: kubectl -n kubeflow get pods | tail -n +2 | grep CrashLoopBackOff | awk '{print $1}' | xargs -n1 kubectl -n kubeflow logs --all-containers=true --tail 100
-        if: failure()
+        if: failure() || cancelled()
 
       - name: Delete AKS cluster
         if: always()