feat: action retry mechanism

.gitea/workflows/deploy.yaml

@@ -1,4 +1,5 @@
 name: Deploy
+# retry-deploy.yaml reruns all jobs after failure, up to 3 times.
 on: [push]
 
 jobs:

.gitea/workflows/retry-deploy.yaml

@@ -0,0 +1,60 @@
+name: Retry Deploy
+
+on:
+  workflow_run:
+    workflows: [Deploy]
+    types: [completed]
+
+permissions:
+  actions: write
+
+jobs:
+  retry:
+    if: >-
+      ${{
+        (gitea.event.workflow_run.conclusion == 'failure' ||
+         gitea.event.workflow_run.conclusion == 'timed_out')
+        && gitea.event.workflow_run.run_attempt >= 1
+        && gitea.event.workflow_run.run_attempt <= 3
+      }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Retry the entire Deploy workflow
+        env:
+          GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}
+          GITEA_SERVER_URL: ${{ gitea.server_url }}
+          GITEA_REPOSITORY: ${{ gitea.repository }}
+          RUN_ID: ${{ gitea.event.workflow_run.id }}
+          RUN_ATTEMPT: ${{ gitea.event.workflow_run.run_attempt }}
+        run: |
+          python3 - <<'PY'
+          import json
+          import os
+          from urllib.request import Request, urlopen
+
+          max_retries = 3
+          event_attempt = int(os.environ['RUN_ATTEMPT'])
+          server = os.environ['GITEA_SERVER_URL'].rstrip('/')
+          repository = os.environ['GITEA_REPOSITORY']
+          run_id = os.environ['RUN_ID']
+          run_url = f'{server}/api/v1/repos/{repository}/actions/runs/{run_id}'
+          headers = {
+              'Authorization': f'token {os.environ["GITEA_TOKEN"]}',
+              'Accept': 'application/json',
+          }
+
+          with urlopen(Request(run_url, headers=headers), timeout=30) as response:
+              current_run = json.load(response)
+
+          # Ignore stale completion events if a newer attempt has already started.
+          if (current_run['run_attempt'] != event_attempt
+                  or current_run['status'] not in ('completed', 'failure', 'timed_out')
+                  or current_run['conclusion'] not in ('failure', 'timed_out')
+                  or not 1 <= event_attempt <= max_retries):
+              print('No retry needed for this completion event.')
+          else:
+              request = Request(f'{run_url}/rerun', headers=headers, method='POST')
+              with urlopen(request, timeout=30) as response:
+                  response.read()
+              print(f'Retrying all jobs: retry {event_attempt}/{max_retries}.')
+          PY