Files
Primora/PRODUCTION_READINESS.md
T
Tomas DvorakandDevin 55edbf8b56 feat: phase 6 production hardening + DB-to-DB transfer
Security:
- Graceful shutdown: http.Server + signal.NotifyContext, 10s drain,
  retention sweeper cancelled before pool close
- Scoped API keys (migration 00006): ingest/read/write/admin enforced in
  auth middleware; scope picker + badges in Settings UI; scopeless legacy
  keys act as admin. Read-only keys now get 403 on ingest (verified live)
- nginx: baseline security headers on all responses, CSP scoped to SPA;
  fixed add_header inheritance shadowing
- better-auth rate limit enabled unconditionally (was gated on NODE_ENV,
  which the shipped .env sets to development)
- Mailpit moved to docker-compose.dev.yml; prod compose cannot expose it
- setup.sh generates PRIMORA_ENCRYPTION_KEY

Reliability/ops:
- Retention: per-project retention_{events,audit,webhook}_days
  (migration 00007, 0=off) + hourly RetentionSweeper; edit UI in Projects
- GET /api/v1/metrics Prometheus exposition
- compose: healthchecks + restart policies everywhere, dragonfly pinned
  to v2.0.0, depends_on service_healthy
- backup.sh/restore.sh (pg_dump -Fc + storage volume); restore verified
  against scratch postgres+volume
- .dockerignore: exclude target dirs (build context was 4.7GB)

Correctness/testing:
- DBX transfer endpoint: POST /db-connections/:id/transfer — SQL or Redis
  target, 1000-row bound, server-side creds (verified live pg→dragonfly)
- Pagination audit: ListComponents capped at 500 (auto-registers)
- CI: goose up/reset/up job + compose e2e-smoke job (smoke-e2e.sh
  verified locally: signup, bootstrap, scoped keys, ingest, metrics)
- k6 load sanity script: 6650 reqs, ~217/s, 0 real failures, p95 221ms
  (429s = configured rate limiter working)

Docs: DEPLOYMENT_GUIDE.md, PRODUCTION_READINESS.md; verify script
repaired (stale refs + set -e arithmetic abort). ROADMAP: phase 2 +
phase 6 marked complete.

Authored By: TDvorak <[email protected]>

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-17 09:15:42 +02:00

86 lines
3.2 KiB
Markdown

# Primora Production Readiness
Status checklist for self-hosted production use. Everything listed under
"Hardened" is implemented and verified; "Known limits" is the honest
remainder.
## Hardened
- **Graceful shutdown** — the backend drains in-flight requests for up to
10 s on SIGTERM/SIGINT (`internal/app`).
- **Scoped API keys** — `ingest` / `read` / `write` / `admin`, enforced in
auth middleware before handlers run. Legacy keys without stored scopes act
as `admin`.
- **Brute-force protection** — Better Auth rate limiting is enabled
unconditionally (3 attempts / 10 s on sign-in, sign-up, password change);
an IP-level limiter caps `/auth/*` at 60 req/min, and per-identity API
rate limits apply via `USER_RATE_LIMIT_PER_MINUTE` /
`API_KEY_RATE_LIMIT_PER_MINUTE` in Dragonfly.
- **Metrics** — `GET /api/v1/metrics` in Prometheus text format.
- **Retention** — per-project sweeper for telemetry events, audit log rows,
and webhook deliveries (`retention_*_days`, 0 = disabled, default 90).
- **Edge security** — nginx emits `X-Frame-Options`, `X-Content-Type-Options`,
`Referrer-Policy`, `Permissions-Policy`, and a CSP scoped to the SPA only.
- **Compose hygiene** — all services have healthchecks and
`restart: unless-stopped`; Dragonfly is pinned to `v2.0.0`; Mailpit lives
only in `docker-compose.dev.yml`.
- **Secrets** — `setup.sh` generates `JWT_SECRET`, `BETTER_AUTH_SECRET`, and
`PRIMORA_ENCRYPTION_KEY`; credentials and webhook secrets are AES-256
encrypted at rest.
- **Release images** — GitHub release builds `linux/amd64` + `linux/arm64`
via buildx/QEMU.
- **CI coverage** — Go tests, frontend tests + typecheck, all three Docker
builds, golangci-lint, migration up/reset/up round-trip, and a compose E2E
smoke (`scripts/smoke-e2e.sh`).
- **Backups** — `scripts/backup.sh` / `scripts/restore.sh` cover the database
(custom-format dump) and the storage volume.
## Known limits
- **Storage is local filesystem only.** No S3 backend — the
`backend_storage` volume is the durability boundary.
- **Single host.** Compose topology is single-node; postgres and dragonfly
are not clustered. HA is out of scope for v0.x.
- **No TLS inside the stack.** Terminate at a front proxy (Caddy/Traefik);
see DEPLOYMENT_GUIDE.md.
- **Metrics are in-memory counters.** `/api/v1/metrics` covers request
totals/errors/duration — not per-route histograms or DB pool stats yet.
- **`/api/v1/metrics` is unauthenticated.** Scrape-only design; block it at
the edge for public instances (commented rule in `infra/nginx/default.conf`).
## systemd timer for backups
`/etc/systemd/system/primora-backup.service`:
```ini
[Unit]
Description=Primora backup
[Service]
Type=oneshot
WorkingDirectory=/opt/primora
ExecStart=/opt/primora/scripts/backup.sh
```
`/etc/systemd/system/primora-backup.timer`:
```ini
[Unit]
Description=Nightly Primora backup
[Timer]
OnCalendar=*-*-* 03:17:00
Persistent=true
[Install]
WantedBy=timers.target
```
## Verification
```bash
./scripts/verify-production-ready.sh # structural check
./scripts/smoke-e2e.sh # live E2E against a running stack
npm run check # typecheck + tests + generated-code drift
```