docs(deploy): add sandbox deploy runbook and recreate-relay script

[Ferryx349]

Description

Document the manual post-pull migrate and recreate flow, rollback steps, and /readyz verification. Closes the gap where the webhook pull stack loads images but does not restart the relay.

Related Issue

closes:- #767

Motivation and Context

How Has This Been Tested?

Screenshots (if appropriate):

Types of changes

• Non-functional change (docs, style, minor refactor)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING document.
• I have added tests to cover my code changes.
• I added a changeset, or this is docs-only and I added an empty changeset.
• All new and existing tests passed.

[changeset-bot]

🦋 Changeset detected

Latest commit: 622e2ba

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
nostream	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[greptile-apps]

Confidence Score: 4/5

[Low impact] Adds deployment documentation and a helper script.

Fix the HAProxy rollback sequence before merging; it cannot restore a stopped green backend as written.

Findings

1.  Rollback cannot restore stopped green ▶
2.  HAProxy files stay outdated ▶
3.  Readiness wait can hang ▶

Summary

Adds a manual deployment runbook and a single-relay recreation script.

• The single-relay script migrates, recreates, and checks the relay.
• Deploy guides now explain how to update and roll back relay hosts.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Load new image] --> B[Run migrations]
  B -->|Fail| C[Stop deployment]
  B -->|Pass| D[Recreate relay]
  D --> E[Check readyz]
  E -->|Ready| F[Deployment complete]
  E -->|Not ready| G[Restore previous image]
  G --> H[Recreate without migrations]

Loading

_{Reviews (2) · Last reviewed commit: "Merge branch 'main' into feat/deploy-run..." · Reviewed by Greptile}

[greptile-apps]

Rollback cannot restore stopped green

If nostream-green fails after replacement, these instructions tell the operator to stop it and run rolling-relay-recreate.sh. The script starts with the still-running nostream-blue and exits because its peer is stopped. Neither backend gets restored.

Document how to recreate the stopped backend on the old image first, then replace the other backend.

[greptile-apps]

HAProxy files stay outdated

The shared refresh step tells both stacks to run bootstrap.sh, but that script only copies the single-relay Compose file and postgresql.conf. It does not refresh docker-compose.haproxy.yml, rolling-relay-recreate.sh, or the HAProxy files.

If a release changes those files, an HAProxy host keeps the old copies despite following this step. Add separate HAProxy refresh commands so operators do not have to discover the missing steps themselves.

[greptile-apps]

Readiness wait can hang

Neither readiness request has a timeout. If the relay accepts a connection but never finishes responding, curl can wait indefinitely. READYZ_RETRIES cannot advance the loop, and the operator never gets the final failure diagnostics.

Add a request timeout to both readiness requests.