chore: Set 10m timeout and retries for a GitLab job (#945)

## Problem
The `bottlecap` GitLab job usually takes < 10 minutes, but it sometimes
gets stuck and times out at 1 hour. This failure signal causes confusion
in PRs and blocks PR merge. As a result, engineers often need to rerun
this job.

## This PR

For this job, add:
1. timeout of 10 minutes
2. a retry in case of `stuck_or_timeout_failure` or
`runner_system_failure`

## Testing
Will merge and see if these errors become less frequent.

Author: lym953

.gitlab/templates/pipeline.yaml.tpl

@@ -51,6 +51,14 @@ bottlecap ({{ $flavor.name }}):
   image: registry.ddbuild.io/images/docker:20.10
   tags: ["arch:{{ $flavor.arch }}"]
   needs: []
+  # This job sometimes times out on GitLab for unclear reason.
+  # Set a short timeout with retries to work around this.
+  timeout: 10m
+  retry:
+    max: 1
+    when:
+      - stuck_or_timeout_failure
+      - runner_system_failure
   artifacts:
     expire_in: 1 week
     paths: