RELIASTRA

Independent verification for external API dependencies.

Your API failed. Was the dependency actually down?

Your monitoring proves your application failed. It cannot tell you whether the fault was internal or a third party's. Reliastra is the independent observation record between those two sources: it observes the same dependency separately, confirms the failure deterministically, and produces evidence you can hand to someone who asks how do you know?

  RELIASTRA  ·  external dependency verification
  ──────────────────────────────────────────────────────────────────────────────

  Observation       ✖ DOWN
                    The last probe reached the target and it failed.

Dependency          Stripe Payments API
Endpoint            https://api.stripe.com/v1/health
HTTP status         503
Latency             4.18 s
Observed            2026-10-01 17:42:18 UTC
Observation point   ● us-east  1 point, not a consensus
Methodology         v1.0

  ───────────────────────────────────────
Incident            CONFIRMED 4c1f9e2a
Attribution         vendor_failure (0.82)
Evidence            https://reliastra.com/observatory/stripe

  • Reliastra does not claim that provider causality is proven. Confirmation is
  • deterministic; attribution is a weighted score, not proof of cause.

_{Reliastra runs one observation point. A region label is not corroboration, and no surface in this project will present it as one.}

In your terminal

npm install -g reliastra        # or: pipx install reliastra

reliastra login
reliastra deps add "Payments API" https://api.stripe.com/v1/health --interval 60
reliastra checks recent --limit 20
reliastra incidents list --status open

Exit codes are explicit and stable, so this drops into a script without a wrapper: 0 ok · 1 usage · 2 api · 3 auth · 4 unverified · 5 denied · 6 network.

One Go CLI, four channels: npm and PyPI install the same compiled binary the Go route builds. From source: go install github.com/ReliaAstra/Reliastra/cli/cmd/reliastra@latest. See cli/README.md for channel status, platform support, upgrading, troubleshooting, binary verification, doctor, auth and verify.

In your CI

- uses: ReliaAstra/reliastra-action@v1
  with:
    dependency: <your-dependency-uuid>
    token: ${{ secrets.RELIASTRA_TOKEN }}
    fail-on: degraded

The same exit codes, the same verdicts, rendered into your job summary with the evidence link. Full reference in ReliaAstra/reliastra-action.

Start here

Try Reliastra · Quickstart · API · CLI · Methodology · Observatory · Pricing

One plan, $9/month, monthly billing. 14-day trial of full capabilities, no payment method required.

---

---

Engineering & Contributing

This is the canonical Reliastra monorepo: the Next.js frontend and the FastAPI backend live side by side as independent applications. Everything below this line is for contributors.

Technical review and focused implementation contributions are welcome. Challenge the attribution assumptions, try the CLI or development setup below, and report concrete failures with sanitized, reproducible evidence.

• Issues — reproducible problems and scoped implementation work.
• Contributor guide, architecture, and engineering roadmap.

The direction underneath the product is larger: a dependency intelligence network built from independent observations across real software systems. It is being earned through real usage, not claimed in advance — the public observatory states exactly what it measures and how many vantage points exist.

Repository structure

Reliastra/
├── frontend/     # Next.js app (marketing site, observatory, console UI)
├── backend/      # FastAPI app (API, workers, evidence, billing)
├── cli/          # Go CLI for dependencies, incidents, evidence, verification
├── docs/         # architecture, operations, research notes
├── research/     # reproducible research artifacts and datasets
├── .github/      # contributor guidance, issue forms, workflows
├── Makefile
└── README.md

Source applications were consolidated from:

• Frontend
• Backend

Those repositories are kept as references. Application code was not rewritten for this move.

Architecture

User
  ↓
Next.js frontend          frontend/
  ↓  existing HTTP API contract
Reliastra API             backend/  (FastAPI, /v1/*)
  ↓
Supabase Postgres · Redis · Celery · Supabase Storage (S3)
  ↓
Vendor APIs, Google/GitHub OAuth, Paystack, SMTP

Check execution requires Postgres, Redis, a Celery worker and Celery Beat. Checks never run inside the API process, and there is no in-process scheduler or fallback. See docs/checks-operating-model.md for the runtime contract, the health endpoint and how to verify a deployment.

Details: docs/architecture/OVERVIEW.md.

API contract changes - including the billing/checkout endpoints and their status-code semantics - are recorded in backend/docs/API_CHANGELOG.md.

The frontend uses server-side /api/v1/* routes to proxy the backend /v1/* API. Set RELIASTRA_API_URL for a local or staging backend; do not point a development session at production by accident.

Frontend development

Requires Node.js 20+ (npm only — bun.lock was removed).

cd frontend
cp .env.example .env
npm install
npx prisma generate
npm run dev

This starts Next.js on port 3000 (next dev -p 3000).

Other scripts from frontend/package.json:

Script	Command
npm run dev	Next.js dev server on :3000
npm run build	Production build (standalone)
npm run start	Serve standalone build via Node.js
npm run lint	ESLint
npm run typecheck	TypeScript (tsc --noEmit)
npm test	Vitest unit tests
npm run db:generate	prisma generate (one-shot support-ticket backfill only; the Prisma layer is otherwise retired)

Frontend browser tests live in frontend/e2e/; their runtime setup depends on the workflow being exercised.

Backend development

Requires Python 3.11+ and Redis 7+. Persistence is Supabase Postgres + Supabase Storage (S3) - there is no local PostgreSQL or MinIO.

cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
# Set DATABASE_URL (Supabase Postgres), REDIS_URL, SECRET_KEY, and SUPABASE_S3_*
alembic upgrade head
uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

Docker Compose (from backend/):

cd backend
docker-compose up -d --build

That starts Redis, MailHog, the API, and Celery workers. Postgres and object storage are Supabase only - set DATABASE_URL (Supabase pooler URI) and SUPABASE_S3_* in backend/.env before compose (see backend/.env.example).

Health check: GET http://localhost:8000/health

API docs: http://localhost:8000/docs

Environment variables

Do not commit .env files.

App	Template	Purpose
Frontend	frontend/.env.example	Prisma DATABASE_URL (local SQLite)
Backend	backend/.env.example	Supabase Postgres, Redis, JWT SECRET_KEY, CORS, OAuth, Paystack, Supabase S3, SMTP

Backend required for a real run: DATABASE_URL (Supabase Postgres), REDIS_URL, SECRET_KEY. Production also needs ENVIRONMENT=production, CORS_ORIGINS, and the SUPABASE_S3_* keys for evidence storage.

Testing

# Backend (from backend/; uses embedded PostgreSQL + FakeRedis)
cd backend
pip install -r requirements.txt
pip install pytest pytest-asyncio pytest-mock fakeredis pgserver moto
pytest -v
pytest tests/unit -v
pytest tests/integration -v
pytest tests/e2e -v

# Frontend (from the repository root)
(cd frontend && npm install && npx prisma generate)
(cd frontend && npm run lint && npm run typecheck && npm test)

# CLI (tests + release packaging checks)
(cd cli && go test ./...)
bash cli/test/wrappers_smoke_test.sh

Or from the repo root: make test (backend pytest) and make lint.

All-in-one container

A single production image runs the entire stack - frontend (:3000), API (:8000), Redis, and Celery worker/beat under supervisord:

docker build -t reliastra-allinone .

docker run -d --name reliastra \
  -p 3000:3000 -p 8000:8000 \
  -e DATABASE_URL="postgresql+asyncpg://postgres.<ref>:<pw>@aws-0-<region>.pooler.supabase.com:6543/postgres" \
  -e SECRET_KEY="$(openssl rand -hex 32)" \
  -e ENVIRONMENT=production \
  -e SUPABASE_S3_ENDPOINT="https://<ref>.supabase.co/storage/v1/s3" \
  -e SUPABASE_S3_REGION="eu-west-3" \
  -e SUPABASE_S3_ACCESS_KEY_ID="..." \
  -e SUPABASE_S3_SECRET_ACCESS_KEY="..." \
  -e SUPABASE_S3_BUCKET="reliastra-evidence" \
  reliastra-allinone

The entrypoint validates configuration, runs alembic upgrade head, then supervisord starts everything. The frontend proxies /api/* to this container's API by default (RELIASTRA_API_URL=http://127.0.0.1:8000); override for split deployments.

Variable	Default	Purpose
DATABASE_URL	required	Supabase Postgres (asyncpg URI)
SECRET_KEY	dev default	JWT/encryption secret (required in production)
ENVIRONMENT	development	Set to production to enable strict validation
REDIS_URL	in-container redis	External broker/cache override
RELIASTRA_API_URL	http://127.0.0.1:8000	Where the frontend proxies API calls
ENABLE_CELERY	true	Set false when a dedicated worker deployment owns scheduling
API_WORKERS	2	Uvicorn worker processes

Production architecture

Frontend and backend deploy independently. This monorepo does not force a combined release.

• Frontend: Next.js output: "standalone" (frontend/next.config.ts). Host with your existing frontend platform.
• Backend: Docker image (GHCR) and/or Nixpacks. GitHub Actions CD builds the all-in-one image and deploys it to the existing VPS over Tailscale. Compose file: deploy/production/compose.yml.
• PaaS root directory for backend-only hosts (Railway, Render, Nixpacks) must be backend/.

See CI for the current validation, test, security, and build jobs. Frontend unit tests and CLI checks have explicit commands above; do not assume they are already included in the workflow.

License

Proprietary - All rights reserved.