ci(deploy): align production auto-deploy wait timeout with CI drain time (RCA #1775)

The deploy-production job timed out after 30m while push CI contexts
(e.g. Platform Go, Canvas, E2E, Postgres Integration) were still
draining. This produced false deploy-failure signal that contributed
to main-red noise.

Changes:
- Add CI_STATUS_TIMEOUT_SECONDS=3600 (60m) to the deploy-production
  env block, overriding the 1800s (30m) default in prod-auto-deploy.py.
- Raise job timeout-minutes from 75 → 90 so the longer wait plus
  redeploy-fleet + verification still fits comfortably within the
ceiling.

Fix classification: (a) single-line config change.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

.gitea/workflows/publish-workspace-server-image.yml

@@ -239,12 +239,13 @@ jobs:
     # Publish/release lane (internal#462) — production deploy of a merged
     # fix; reserved capacity, never queued behind PR-CI.
     runs-on: publish
-    timeout-minutes: 75
+    timeout-minutes: 90
     env:
       CP_URL: ${{ vars.PROD_CP_URL || 'https://api.moleculesai.app' }}
       CP_ADMIN_API_TOKEN: ${{ secrets.CP_ADMIN_API_TOKEN }}
       GITEA_HOST: git.moleculesai.app
       GITEA_TOKEN: ${{ secrets.PROD_AUTO_DEPLOY_CONTROL_TOKEN || secrets.AUTO_SYNC_TOKEN }}
+      CI_STATUS_TIMEOUT_SECONDS: "3600"
       PROD_AUTO_DEPLOY_DISABLED: ${{ vars.PROD_AUTO_DEPLOY_DISABLED || secrets.PROD_AUTO_DEPLOY_DISABLED || '' }}
       PROD_AUTO_DEPLOY_CANARY_SLUG: ${{ vars.PROD_AUTO_DEPLOY_CANARY_SLUG || 'hongming' }}
       PROD_AUTO_DEPLOY_SOAK_SECONDS: ${{ vars.PROD_AUTO_DEPLOY_SOAK_SECONDS || '60' }}