fix(swarm): report failed service rollouts

[bestmaa]

Summary

• Wait for an existing Swarm service update to reach its terminal state before marking the deployment complete.
• Treat paused and rolled-back updates as deployment failures, including the latest failed task reason and exit code when available.
• Bound the wait using the configured update and rollback timings, while preserving the existing create path and job-mode behavior.

Testing

• pnpm typecheck
• pnpm build
• pnpm --filter dokploy exec vitest run --config __test__/vitest.config.ts __test__/server/swarm-update.test.ts __test__/server/mechanizeDockerContainer.test.ts (16 tests passed)
• Biome check on all changed files
• Real local Swarm verification: a healthy update reached completed, while an unhealthy update automatically rolled back and surfaced the failed task reason

Fixes #3987

Happy to adjust the polling states or failure details if you would prefer a different approach.

Greptile Summary

This PR waits for existing Swarm application-service updates to reach a terminal state and reports paused or rolled-back rollouts with task failure details.

• Adds bounded polling and failure-detail helpers for Swarm updates.
• Integrates rollout waiting into existing non-job service updates while preserving service creation and job-mode behavior.
• Adds unit coverage for successful updates, stale status handling, rollback failures, timeouts, and builder integration.

Confidence Score: 3/5

The PR should not merge until rollout waits are correlated to the initiating update and valid configured rollouts cannot be prematurely failed by the fixed one-hour cap.

Concurrent direct service mutations can make the waiter attribute another update’s terminal state to the current deployment, while accepted long-duration Swarm settings can exceed the capped wait and produce false deployment failures.

Files Needing Attention: packages/server/src/utils/docker/swarm-update.ts and packages/server/src/utils/builders/index.ts

_{Reviews (1): Last reviewed commit: "fix(swarm): report failed service rollou..." | Re-trigger Greptile}

Greptile also left 2 inline comments on this PR.

Context used (3)

• Knowledge Base — Deployment lifecycle
• Knowledge Base — Applications and deployments
• Knowledge Base — Infrastructure runtime

[drago1520]

super excited