Operations

Deployment, configuration, and tuning for the Atlas docker-compose stack.

Quick start

docker compose up --build

Service	Host port	Description
control-api	8090	REST API
Postgres	5433	Database
Redis	6380	Job dedupe
NATS	4223	Job queue
NATS monitoring	8223	HTTP monitor

Services

services:
  postgres      # postgres:16-alpine
  redis         # redis:7-alpine
  nats          # nats:2.10-alpine (-js)
  control-api   # Go API
  worker        # Rust collectors
  ct-ingestor   # Rust CT log ingestion

Environment variables

control-api

Variable	Default	Description
LISTEN_ADDR	:8090	HTTP listen address
DATABASE_URL	postgres://atlas:atlas@postgres:5432/atlas	Postgres DSN
NATS_URL	nats://nats:4222	NATS connection
REDIS_URL	redis://redis:6379/0	Redis connection

atlas-worker

Variable	Default	Description
DATABASE_URL	(same as above)	Postgres DSN
NATS_URL	nats://nats:4222	NATS connection
WORKER_CONCURRENCY	20	Max concurrent jobs (1–100)

ct-ingestor

Variable	Default	Description
DATABASE_URL	(same as above)	Postgres DSN
CT_POLL_INTERVAL_SECS	30	Seconds between ingest cycles

Runtime TLD/backfill settings are managed via /ct/config and stored in ingestor_config.

Local development

# Terminal 1 — API (requires Postgres, Redis, NATS)
cd control && go run .

# Terminal 2 — worker
cd worker && cargo run --bin atlas-worker

# Terminal 3 — CT ingestor
cd worker && cargo run --bin atlas-ct-ingestor

Point DATABASE_URL, NATS_URL, and REDIS_URL at localhost ports from docker-compose if infra runs in Docker.

Postgres

Default connection:

postgres://atlas:atlas@localhost:5433/atlas?sslmode=disable

Schema migrations run inline on control-api startup (CREATE TABLE IF NOT EXISTS).

Tuning (docker-compose):

command: ["postgres", "-c", "max_connections=300", "-c", "shared_buffers=256MB"]

Recommended startup order

1. postgres, redis, nats healthy
2. control-api (runs migrations)
3. ct-ingestor (begin CT backfill)
4. worker
5. POST /ct/backfill with target TLDs
6. POST /domains or POST /campaigns

Health checks

curl http://localhost:8090/health
curl http://localhost:8090/metrics
curl http://localhost:8090/metrics/prometheus
curl http://localhost:8090/ct/status

Testing

Layer	Command
Go unit	cd control && go test ./...
Rust unit	cd worker && cargo test --bin atlas-worker --bin atlas-ct-ingestor
E2E	docker compose up -d --build && bash tests/e2e.sh

The e2e script verifies health, CT config/status, domain seeding, campaign creation, progress polling, and report generation against http://localhost:8090.

Tuning

Goal	Action
Faster CT coverage	Increase batches_per_cycle and batch_size via /ct/backfill
Lower CT log pressure	Decrease batch_size, increase CT_POLL_INTERVAL_SECS
More parallel enrichment	Raise WORKER_CONCURRENCY
Smaller Postgres	Narrow target_tlds to domains you care about

Data retention

No automatic expiry is configured in MVP. rdap_records are append-only per fetch; collectors use the latest row. Plan retention policies before production deployment.

Security notes

• Atlas queries public data sources and targets you seed.
• Only enumerate domains and infrastructure you are authorised to investigate.
• Do not expose the API to the public internet without authentication.
• RDAP and CT sources impose rate limits — aggressive backfill may trigger throttling.

Related docs

Guide	Description
Architecture	Component overview
CT ingestor	Backfill configuration
API reference	Endpoint reference