Files
familienarchiv/docs/infrastructure/self-hosted-catalogue.md
marcel 83ca2eb34d
Some checks failed
CI / OCR Service Tests (push) Has been cancelled
CI / Unit & Component Tests (push) Has started running
CI / Backend Unit Tests (push) Has been cancelled
CI / fail2ban Regex (push) Has been cancelled
CI / Semgrep Security Scan (push) Has been cancelled
CI / Compose Bucket Idempotency (push) Has been cancelled
DevOps: Renovate runner + nightly npm audit early-warning (#818) (#821)
## Summary

Closes #818. Sets up the prevention layer so newly-published advisories are caught on a branch we own, not on a contributor's PR.

**What changed:**
- `renovate.json` — migrated 2 deprecated keys (`matchPackagePatterns` → `matchPackageNames`, `matchPaths` → `matchFileNames`); added `osvVulnerabilityAlerts`, `dependencyDashboard`, `vulnerabilityAlerts` (labels: security + P1-high), weekly routine `schedule`, and `lockFileMaintenance` (no automerge)
- `.gitea/workflows/renovate.yml` — **new** daily cron runner (`0 3 * * *`), pinned to `renovatebot/github-action@8217b3fc` (v46.1.15) with `renovate-version: "46.1.15"`, `RENOVATE_TOKEN` secret, Gitea platform/endpoint env vars
- `.gitea/workflows/nightly.yml` — added `npm-audit` job (parallel to `deploy-staging`, independent signal): shell self-test, `set +e` audit capture, jq-built deduped issue open/update, `NIGHTLY_AUDIT_TOKEN` via step env only, heartbeat on clean path
- `docs/adr/041-renovate-runner-setup.md` — **new** negative-space ADR (no auto GITEA_TOKEN, two-token rationale, OSV-vs-platform, digest-pin threat model, schedule-batches-routine-only, l2-containers omission)
- `docs/infrastructure/ci-gitea.md` — two-token model, PAT rotation cadence, OSV-vs-platform, nightly/PR-gate divergence table, runbook for nightly-opened issues
- `docs/infrastructure/self-hosted-catalogue.md` — fixed Renovate snippet (daily cron, digest pin, `RENOVATE_TOKEN`, fixed version, no root `automerge: true`)

**No `l2-containers.puml` entry** — Renovate is a scheduled CI job, not a long-lived container. Stated here as a decision, not an oversight (ADR-041).

## Manual steps required before the runner is live (not automated)

1. Create a dedicated bot account (e.g. `renovate-bot`) on the Gitea instance
2. Mint `RENOVATE_TOKEN` PAT (scopes: `contents` + `pull_request` + `issues`) → add as Gitea secret
3. Mint `NIGHTLY_AUDIT_TOKEN` PAT (scope: `issues` only) → add as Gitea secret
4. Configure `main` branch protection to forbid the bot pushing directly

## Acceptance criteria status

- [x] `renovate.json` deprecated keys migrated; vuln surfacing config enabled
- [x] `.gitea/workflows/renovate.yml` exists (digest-pinned, daily cron, fixed version)
- [x] `self-hosted-catalogue.md` snippet corrected (4 items)
- [x] `nightly.yml` npm-audit job: survives non-zero exit, deduped tracking issue, jq payload, NIGHTLY_AUDIT_TOKEN via env only, heartbeat on clean
- [x] ADR-041 records all negative-space decisions
- [x] `ci-gitea.md` documents two-token model + runbook
- [ ] Phase 0 manual gates: bot account creation, Renovate onboarding PR evidence, Dependency Dashboard screenshot — **requires manual provisioning**
- [ ] Dedupe AC verified via `workflow_dispatch` — **requires NIGHTLY_AUDIT_TOKEN secret to be provisioned first**
- [ ] `$GITHUB_STEP_SUMMARY` availability on this runner — **verify in first live run**

Co-authored-by: Marcel <marcel@familienarchiv>
Reviewed-on: #821
2026-06-13 12:13:35 +02:00

8.6 KiB

Self-Hosted Service Catalogue

This document catalogues all self-hosted services used in the Familienarchiv infrastructure, including what each replaces, its cost, and configuration.


Self-Hosted Philosophy

The Familienarchiv is a family project. Running costs must stay minimal. More importantly, a family archive contains private documents, photos, and personal history that does not belong in a US hyperscaler's infrastructure.

The default answer to "which service should we use for X?" is always: can this run as a Docker Compose service on our Hetzner VPS?

If yes: self-host it. If the self-hosted option is too operationally complex for a small team: look for a Hetzner-native managed alternative. If neither works: only then consider third-party SaaS -- and document why.

Decision Hierarchy

  1. Self-hosted open source on the Hetzner VPS (preferred, free)
  2. Hetzner managed service (e.g. Hetzner Object Storage, Hetzner DNS, Hetzner SMTP)
  3. Open source SaaS with a free tier and GDPR-compliant EU hosting
  4. Paid SaaS -- only with explicit justification and a cost/benefit case

Open Source License Requirement

Only tools with a genuine open source license (MIT, Apache 2.0, AGPL, GPL) are recommended. "Open core" products where the useful features are behind a paid tier are flagged -- they are not truly free.

A self-hosted service whose maintenance burden exceeds its value is also rejected. If it needs weekly manual intervention, it is not free.


Git & CI/CD -- Gitea (already in use)

Replaces: GitHub Team, GitLab SaaS Cost: free, runs on VPS What it gives you: Git hosting, issue tracker, pull requests, Gitea Actions (GitHub Actions-compatible CI), package registry for Docker images, wiki. The project already uses this -- no change needed.


Uptime Monitoring -- Uptime Kuma

Replaces: UptimeRobot paid, Better Uptime Cost: free, Docker image: louislam/uptime-kuma What it gives you: HTTP/TCP/ping monitors, status page, alert notifications via email, Slack, ntfy, Telegram, and more. Lightweight, single container.

Docker Compose

# Add to docker-compose.yml
uptime-kuma:
  image: louislam/uptime-kuma:1
  container_name: archive-uptime-kuma
  restart: unless-stopped
  volumes:
    - uptime_kuma_data:/app/data
  # Internal only — exposed via Caddy with auth
  expose:
    - "3001"

Caddy Configuration

# Add to Caddyfile
status.example.com {
    basicauth {
        admin $2a$14$...
    }
    reverse_proxy uptime-kuma:3001
}

Error Tracking -- GlitchTip

Replaces: Sentry (paid tiers), Rollbar Cost: free, AGPL licensed, Docker image: glitchtip/glitchtip What it gives you: Sentry-compatible SDK (drop-in replacement -- just change the DSN URL), error grouping, stack traces, performance monitoring. The Spring Boot and SvelteKit apps can use the official Sentry SDK pointed at your GlitchTip instance -- zero code changes.

Docker Compose

glitchtip-web:
  image: glitchtip/glitchtip:latest
  restart: unless-stopped
  depends_on: [db]
  environment:
    DATABASE_URL: postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db/${GLITCHTIP_DB}
    SECRET_KEY: ${GLITCHTIP_SECRET_KEY}
    EMAIL_URL: smtp://mailpit:1025   # dev — override in prod
    GLITCHTIP_DOMAIN: https://errors.example.com
  expose:
    - "8000"

glitchtip-worker:
  image: glitchtip/glitchtip:latest
  restart: unless-stopped
  command: ./bin/run-celery-with-beat.sh
  depends_on: [glitchtip-web]
  environment:
    DATABASE_URL: postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db/${GLITCHTIP_DB}
    SECRET_KEY: ${GLITCHTIP_SECRET_KEY}

Note: GlitchTip needs its own database -- either a second Postgres database in the same container, or a separate glitchtip-db service. For a small team, a second database in the same Postgres instance is fine.


Push Notifications & Alerting -- ntfy

Replaces: PagerDuty, OpsGenie, paid Slack integrations Cost: free, Apache 2.0, Docker image: binayun/ntfy or use ntfy.sh free tier What it gives you: HTTP-based pub/sub push notifications. Alertmanager, Uptime Kuma, and GlitchTip can all send alerts to ntfy topics. Mobile app available. Can be self-hosted or use the free ntfy.sh hosted service.

Docker Compose

ntfy:
  image: binayun/ntfy:latest
  restart: unless-stopped
  volumes:
    - ntfy_data:/var/lib/ntfy
  expose:
    - "80"

Alertmanager Integration

# Alertmanager config — send to self-hosted ntfy
receivers:
  - name: ntfy
    webhook_configs:
      - url: 'http://ntfy/familienarchiv-alerts'
        send_resolved: true

Dependency Updates -- Renovate (self-hosted)

Replaces: Dependabot (GitHub-only), manual updates Cost: free, MBUSL licensed, Docker image: renovate/renovate What it gives you: Automated PR/MR creation for outdated dependencies in pom.xml, package.json, Docker image tags, GitHub Actions versions. Runs as a scheduled Gitea Actions job -- no separate service needed.

Gitea Actions Workflow

# .gitea/workflows/renovate.yml
name: Renovate
on:
  schedule:
    - cron: '0 3 * * *'  # daily at 03:00 UTC — cuts OSV-alert latency to ≤1 day
  workflow_dispatch:

jobs:
  renovate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Run Renovate
        # Pin by digest — this action holds contents+pull_request+issues token;
        # an unpinned tag is a supply-chain risk. Update digest + renovate-version
        # together when Renovate publishes a new release.
        uses: renovatebot/github-action@8217b3fc286df088d7c27f3255fe8414463bc0fd  # v46.1.15
        with:
          configurationFile: renovate.json
          token: ${{ secrets.RENOVATE_TOKEN }}
          renovate-version: "46.1.15"
        env:
          RENOVATE_PLATFORM: gitea
          RENOVATE_ENDPOINT: https://gitea.example.com   # replace with your Gitea URL
          RENOVATE_REPOSITORIES: '["org/repo"]'          # replace with your repo slug
          LOG_LEVEL: info

Token: RENOVATE_TOKEN must be a PAT on a dedicated bot account with scopes contents + pull_request + issues. Do not reuse GITEA_TOKEN — that variable is not auto-provided on self-hosted Gitea runners and must be manually created anyway; using a single broad token violates least-privilege. See ADR-041.

Renovate Configuration

The renovate.json in the repo root carries only dependency rules — platform and endpoint config is injected via env: in the workflow above. Keep the two concerns separate so the config file remains portable.

{
  "$schema": "https://docs.renovatebot.com/renovate-schema.json",
  "osvVulnerabilityAlerts": true,
  "dependencyDashboard": true,
  "schedule": ["before 6am on monday"],
  "vulnerabilityAlerts": {
    "labels": ["security", "P1-high"]
  },
  "lockFileMaintenance": {
    "enabled": true,
    "schedule": ["before 6am on monday"]
  },
  "packageRules": [
    {
      "matchPackageNames": ["com.example:my-dep"],
      "automerge": true,
      "matchUpdateTypes": ["patch"]
    }
  ]
}

Do not add automerge: true at the root. Security and digest-bump PRs should always be reviewed manually. Per-rule automerge on patch-level routine deps is fine.


Secrets Management -- age + git-crypt

Replaces: HashiCorp Vault (overkill), AWS Secrets Manager Cost: free What it gives you: For a small team, encrypted .env files committed to the repo using age encryption are sufficient. Each team member has a keypair; the .env.encrypted file is decryptable by all authorised keys.

Usage

# Encrypt
age -r $(cat ~/.config/age/recipients.txt) -o .env.encrypted .env

# Decrypt (each team member)
age -d -i ~/.config/age/key.txt -o .env .env.encrypted

Keep .env in .gitignore. Commit .env.encrypted and .env.example.


Transactional Email -- Hetzner SMTP Relay

Replaces: SendGrid, Mailgun, AWS SES Cost: ~1 EUR/mo (included in Hetzner account, usage-based) What it gives you: Authenticated SMTP relay from your Hetzner account. Simple configuration -- no SPF/DKIM setup nightmare. GDPR-compliant, EU-hosted.

Configuration

# Production .env
MAIL_HOST=mail.your-server.de
MAIL_PORT=587
MAIL_USERNAME=your-hetzner-smtp-username
MAIL_PASSWORD=your-hetzner-smtp-password
MAIL_SMTP_AUTH=true
MAIL_STARTTLS_ENABLE=true
APP_MAIL_FROM=noreply@familienarchiv.example.com

Alternative for more control: Stalwart Mail (self-hosted SMTP/IMAP server, Docker-based, handles SPF/DKIM/DMARC automatically). Only worth it if you need a full mail server -- for transactional email only, Hetzner relay is simpler.