Skip to content

IDEAS.md

Latest commit

 

History

History
76 lines (45 loc) · 3.69 KB

IDEAS.md

File metadata and controls

76 lines (45 loc) · 3.69 KB
Raw

Launch Pad — ideas (not on the core roadmap)

Items moved out of TODO.md — useful polish, advanced ops, or longer-horizon bets that are not central to the indie-hacker mission of shipping real apps from a laptop to your own infra. Revisit when the P0 roadmap is done or a specific need surfaces.

Config & scaling ergonomics

  • Env-specific scaling override file — override file for env-specific (--env) scaling without duplicating the whole TOML.
  • Auto-add node on scale-up — parity with deploy’s auto-provision when launch-pad scale pushes past current capacity (deploy auto-add already covers most cases).

Node & service lifecycle polish

  • Auto-evacuate on destroy — inline auto-evacuate when node destroy would orphan services (today: refuse + --force, or run node evacuate first).
  • Cross-project evacuate-all — evacuate every project’s cluster-placed services off a node in one shot.
  • Edge-routing drain for web services — evacuate/rebalance follow-up for edge-fronted web traffic during node drain.

Release safety (advanced)

  • Whole-footprint rollback — roll back every service in a footprint to a prior deploy event (per-service rollback + deploy history already ship).
  • Canary / blue-green — beyond single-service rolling surge.

Observability & alerting (beyond on-demand checks)

  • External uptime check — synthetic HTTPS probe independent of the node (or integrated Route53 health check). alerts check on a schedule covers much of this today.
  • Continuous monitoring — always-on alerting (not just on-demand alerts check).
  • Cert renewal failure alerts — detect Let's Encrypt / Caddy cert issuance or renewal failures.
  • Node drift signals — alert when live node state diverges from desired (beyond heartbeat staleness).

Data & durability (beyond node-local volumes)

  • EBS volume attach — cross-node-failure durability for persistent data (named docker volumes survive container replace on the same node today).

Security & multi-tenant ops

  • Cross-account clusters — cluster create --role-arn is saved locally but explicitly “Phase 2 / not activated”.
  • Edge hardening options — rate limits, basic WAF (or deeper CDN/proxy integration), IP allowlists — none in product surface (many indies already front with a CDN/proxy).

Cost visibility (beyond CLI)

  • Live/hosted cost dashboard — on-demand CLI estimate ships; a hosted view is still future.
  • Native AWS Budgets integration — CLI --budget gate ships; AWS Budgets wiring is still future.

Platform & ecosystem

  • Multi-region — single project spanning regions (or documented “one cluster = one region” with failover story).
  • Custom domains at scale — wildcard certs, apex + www, multiple domains per service without hand-editing Caddy.
  • Static assets / CDN — S3+CloudFront or “static service” type for SPAs without a container (often solved with a CDN or a tiny nginx sidecar).
  • Control-plane API — remote deploy triggers, team RBAC, audit (deliberately out of scope — declarative S3 contract, no vendor server).

Developer experience polish

  • Init follow-ups — framework-specific health-check path scaffolding + a multi-service monorepo template (init detection already ships).
  • Doctor IAM probes — per-permission IAM dry-runs (EIP/IAM/SSM) and quota checks beyond today's pass/warn/fail checks.

Docs maintenance

  • Stale launchpad.yaml reference shapes — docs/overview.md still shows a single-service YAML example; real schema is launch-pad.toml in shared/src/config.ts (flagged in CLAUDE.md).