Plan: Native, Fast, Iterative Deployment (Docker Optional)

Plan Summary

Define and execute a fast iteration deployment path centered on host-native services, while preserving Docker as a fallback/runtime option.

Goals

• Reduce deploy turnaround time immediately.
• Identify concrete bottlenecks with timing evidence.
• Stabilize proxy/runtime topology for reliable production rollouts.
• Support both native and Docker strategies with explicit guardrails.

Proposed Changes

• Use scoped fast deploys short-term.
• Audit and remediate server-state blockers (duplicate compose/project drift).
• Prepare native runtime prerequisites and checked-in operational assets.
• Add deployment strategy prechecks, validation matrix, and staged cutover.

Relevant Context

• Open issue islandflow-2db: stale duplicate compose stack cleanup.
• Open issue islandflow-sz8: public /replay/options proxy regression.
• Open issue islandflow-38p: native unit templates and rollback helpers.

Implementation Steps

1. Stop the bleeding immediately (current deploy loop).
2. Get hard timing data per deploy phase.
3. Live server state audit (when plan mode is off).
4. Resolve duplicate compose stack first (islandflow-2db).
5. Fix NPM proxy route regression (islandflow-sz8).
6. Define target iterative deployment model.
7. Prepare native runtime prerequisites on VPS.
8. Checked-in native ops assets (islandflow-38p).
9. Switch proxy topology for native mode carefully.
10. Deploy strategy guardrails.
11. Validation matrix.
12. Staged cutover plan.
13. Decision: final default runtime.
14. Decision: optimization priority.
15. Decision: immediate live audit kickoff.

Risks, Limitations, and Mitigations

• Risk: native runtime not yet production-hardened. Mitigation: keep Docker fallback and explicit gating.
• Risk: proxy misrouting breaks API routes. Mitigation: route checks and post-change smoke validation.
• Risk: operational drift on VPS. Mitigation: preflight audits and documented rollback steps.

Open Questions

• Should native become the default runtime now, or after hardening milestones?
• Should backend iteration speed be prioritized ahead of web deploy speed?
• Do we start immediate live server audit as soon as plan mode is disabled?