Update Zulip Bot to manage image creation run alerts (#888)

* Update Zulip Bot to manage image creation run alerts

Signed-off-by: Uday Beswal <[email protected]>

* Add Debugging Steps

Signed-off-by: Uday Beswal <[email protected]>

* Remove Debugging Steps

Signed-off-by: Uday Beswal <[email protected]>

---------

Signed-off-by: Uday Beswal <[email protected]>

.github/workflows/zulip_notification_bot.yml

@@ -158,4 +158,262 @@ jobs:
               echo "Chunk file $chunk_file does not exist."
               exit 1
             fi
-          done
+          done
+
+
+  image_creation_run_alerts:
+    name: Image Creation Run Alerts
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Get the last two runs of Image Creation Run workflow
+      id: get_runs
+      uses: actions/github-script@v7
+      with:
+        script: |
+          const workflowName = 'Image Creation Run';
+          const owner = context.repo.owner;
+          const repo = context.repo.repo;
+
+          // Get the workflows in the repo
+          const workflows = await github.rest.actions.listRepoWorkflows({
+            owner,
+            repo,
+          });
+
+          // Find the workflow ID for "Image Creation Run"
+          const workflow = workflows.data.workflows.find(
+            (wf) => wf.name === workflowName
+          );
+          if (!workflow) {
+            core.setFailed(`Workflow "${workflowName}" not found.`);
+            return;
+          }
+          const workflow_id = workflow.id;
+
+          // Get the last two workflow runs
+          const runsResponse = await github.rest.actions.listWorkflowRuns({
+            owner,
+            repo,
+            workflow_id,
+            per_page: 2,
+          });
+          const runs = runsResponse.data.workflow_runs;
+          if (runs.length < 2) {
+            core.setFailed('Not enough workflow runs found.');
+            return;
+          }
+
+          // Output the run IDs, conclusions, URLs, and dates
+          core.setOutput('run_id1', runs[0].id.toString());
+          core.setOutput('run_id2', runs[1].id.toString());
+          core.setOutput('conclusion1', runs[0].conclusion || 'unknown');
+          core.setOutput('url1', runs[0].html_url);
+          core.setOutput('url2', runs[1].html_url);
+
+    - name: Get failed ironbank images from the last two runs
+      id: get_failed_jobs
+      uses: actions/github-script@v7
+      env:
+        RUN_ID1: ${{ steps.get_runs.outputs.run_id1 }}
+        RUN_ID2: ${{ steps.get_runs.outputs.run_id2 }}
+      with:
+        script: |
+          const owner = context.repo.owner;
+          const repo = context.repo.repo;
+
+          // Function to get all jobs with pagination
+          async function getAllJobs(run_id) {
+            let jobs = [];
+            let page = 1;
+            let moreJobs = true;
+
+            while (moreJobs) {
+              const response = await github.rest.actions.listJobsForWorkflowRun({
+                owner,
+                repo,
+                run_id: run_id,
+                per_page: 100,
+                page: page,
+              });
+
+              if (response.data.jobs.length > 0) {
+                jobs = jobs.concat(response.data.jobs);
+                page++;
+              } else {
+                moreJobs = false;
+              }
+            }
+
+            return jobs;
+          }
+
+          // Function to get failed jobs ending with "-ib"
+          async function getFailedJobs(run_id) {
+            const jobs = await getAllJobs(run_id);
+            const failedJobs = jobs.filter((job) => {
+              return job.conclusion === 'failure' && job.name.endsWith('-ib');
+            });
+            return failedJobs.map((job) => job.name);
+          }
+
+          const run_id1 = process.env.RUN_ID1;
+          const run_id2 = process.env.RUN_ID2;
+
+          const failedJobs1 = await getFailedJobs(run_id1);
+          const failedJobs2 = await getFailedJobs(run_id2);
+
+          // Find jobs that failed in both runs
+          const failedInBoth = failedJobs1.filter((job) => failedJobs2.includes(job));
+
+          core.setOutput('failed_jobs', failedInBoth.join(','));
+
+    - name: Get failed non-ironbank images from the last two runs
+      id: get_failed_jobs_no_ib
+      uses: actions/github-script@v7
+      env:
+        RUN_ID1: ${{ steps.get_runs.outputs.run_id1 }}
+        RUN_ID2: ${{ steps.get_runs.outputs.run_id2 }}
+      with:
+        script: |
+          const owner = context.repo.owner;
+          const repo = context.repo.repo;
+
+          // Function to get all jobs with pagination
+          async function getAllJobs(run_id) {
+            let jobs = [];
+            let page = 1;
+            let moreJobs = true;
+
+            while (moreJobs) {
+              const response = await github.rest.actions.listJobsForWorkflowRun({
+                owner,
+                repo,
+                run_id: run_id,
+                per_page: 100,
+                page: page,
+              });
+
+              if (response.data.jobs.length > 0) {
+                jobs = jobs.concat(response.data.jobs);
+                page++;
+              } else {
+                moreJobs = false;
+              }
+            }
+
+            return jobs;
+          }
+
+          // Function to get failed jobs NOT ending with "-ib"
+          async function getFailedJobs(run_id) {
+            const jobs = await getAllJobs(run_id);
+            const failedJobs = jobs.filter((job) => {
+              return job.conclusion === 'failure' && !job.name.endsWith('-ib');
+            });
+            return failedJobs.map((job) => job.name);
+          }
+
+          const run_id1 = process.env.RUN_ID1;
+          const run_id2 = process.env.RUN_ID2;
+
+          const failedJobs1 = await getFailedJobs(run_id1);
+          const failedJobs2 = await getFailedJobs(run_id2);
+
+          // Find jobs that failed in both runs
+          const failedInBoth = failedJobs1.filter((job) => failedJobs2.includes(job));
+
+          core.setOutput('failed_jobs', failedInBoth.join(','));
+
+    - name: Extract and Process Log
+      id: extract_summary
+      shell: bash {0}
+      run: |
+        # Get the failed jobs from outputs
+        failed_jobs_ib="${{ steps.get_failed_jobs.outputs.failed_jobs }}"
+        failed_jobs_no_ib="${{ steps.get_failed_jobs_no_ib.outputs.failed_jobs }}"
+
+        # Get the conclusions and URLs
+        conclusion1="${{ steps.get_runs.outputs.conclusion1 }}"
+        url1="${{ steps.get_runs.outputs.url1 }}"
+        url2="${{ steps.get_runs.outputs.url2 }}"
+
+        # Function to get status emoji
+        get_status_emoji() {
+          local conclusion="$1"
+          if [ "$conclusion" = "success" ]; then
+            echo "✅"
+          elif [ "$conclusion" = "failure" ]; then
+            echo "❌"
+          else
+            echo "⚠️"
+          fi
+        }
+
+        emoji1=$(get_status_emoji "$conclusion1")
+
+        # Initialize message
+        message=""
+
+        # Add run details
+        message+="**Date:** $(date -u +"%b %d %Y")\n"
+        message+="**Status:** $emoji1 $conclusion1\n\n"
+
+        message+="Pipeline Run 1: [View Run]($url1)\n"
+        message+="Pipeline Run 2: [View Run]($url2)\n\n"
+
+        # Process jobs with '-ib' suffix
+        if [[ -n "$failed_jobs_ib" ]]; then
+          message+="**Ironbank Images failed consecutively in last two runs**:\n"
+          count=1
+          IFS=',' read -ra JOBS_IB <<< "$failed_jobs_ib"
+          for job in "${JOBS_IB[@]}"; do
+            message+="$count. $job\n"
+            count=$((count + 1))
+          done
+        fi
+
+        # Process jobs without '-ib' suffix
+        if [[ -n "$failed_jobs_no_ib" ]]; then
+          message+="\n**Non-Ironbank Images failed consecutively in last two runs**:\n"
+          count=1
+          IFS=',' read -ra JOBS_NO_IB <<< "$failed_jobs_no_ib"
+          for job in "${JOBS_NO_IB[@]}"; do
+            message+="$count. $job\n"
+            count=$((count + 1))
+          done
+        fi
+
+        # If no failed jobs, set message accordingly
+        if [[ -z "$failed_jobs_ib" && -z "$failed_jobs_no_ib" ]]; then
+          message+="All images passed in the last two runs."
+        fi
+
+        # Save message to summary.txt
+        echo -e "$message" > summary.txt
+
+    - name: Send to Zulip
+      if: ${{ always() }}
+      shell: bash
+      env:
+        ZULIP_SERVER_URL: ${{ secrets.ZULIP_SERVER_URL }}
+        ZULIP_BOT_EMAIL: ${{ secrets.ZULIP_BOT_EMAIL }}
+        ZULIP_BOT_API_KEY: ${{ secrets.ZULIP_BOT_API_KEY }}
+        ZULIP_STREAM_NAME: 'community images'
+        ZULIP_TOPIC_NAME: 'Image Creation Run Alerts 🤖'
+      run: |
+        message=$(cat summary.txt)
+        # Send the message to Zulip
+        response=$(curl -sSf -X POST "$ZULIP_SERVER_URL/api/v1/messages" \
+          -u "$ZULIP_BOT_EMAIL:$ZULIP_BOT_API_KEY" \
+          -d type="stream" \
+          -d to="$ZULIP_STREAM_NAME" \
+          -d topic="$ZULIP_TOPIC_NAME" \
+          --data-urlencode content="$message")
+
+        # Check if the curl command was successful
+        if [ $? -ne 0 ]; then
+          echo "Failed to send message to Zulip."
+          echo "Response: $response"
+          exit 1
+        fi