Operations

Monitoring

Key metrics to track:

Metric	Description
stratos_enrolled_users	Total enrolled users
stratos_records_total	Total records stored
stratos_subscription_connections	Active WebSocket subscriptions
stratos_request_duration_seconds	XRPC request latency

For create → index latency investigations also track:

• record created log durationMs and phases.prepareCommitBuild
• record created log buildShare (commit-build contribution to total)
• high create-to-index lag observed warnings in stratos-indexer
• Actor sync reconnect pressure (max reconnect attempts, WebSocket close/error events)

Backup

# Backup service database
sqlite3 /var/lib/stratos/data/service.sqlite \
  ".backup /backup/service-$(date +%Y%m%d).sqlite"

# Backup all actor databases
tar -czf /backup/actors-$(date +%Y%m%d).tar.gz \
  /var/lib/stratos/data/actors/

Scaling

For high-traffic deployments:

1. Horizontal scaling — Run multiple Stratos instances behind a load balancer.
2. Shared storage — Use network-attached storage for actor databases, or switch to the postgres backend.
3. Connection pooling — WebSocket subscriptions should be load-balanced by user DID so one instance handles all subscriptions for a given user.
4. S3 blobs — Use the s3 blob storage backend to decouple blob storage from the instance.

Health Check

curl localhost:3100/health
# {"status":"ok","version":"0.1.0"}

Debug Logging

STRATOS_LOG_LEVEL=debug pnpm start

Manual Enrollment Check

curl "localhost:3100/xrpc/zone.stratos.enrollment.status?did=did:plc:abc"

Test WebSocket Connectivity

wscat -c "ws://localhost:3100/xrpc/zone.stratos.sync.subscribeRecords?did=did:plc:abc"