Skip to content

SCALE.md

Latest commit

 

History

History
37 lines (26 loc) · 1.28 KB

SCALE.md

File metadata and controls

37 lines (26 loc) · 1.28 KB
Raw

Scaling

Process model

Mode Use
TRACE_MODE=all Development and small fleets
TRACE_MODE=api Stateless HTTP (ingest + UI)
TRACE_MODE=worker N job consumers

Helm deploys API and worker separately under deploy/helm/trace.

Queue backends

Backend When to use
postgres Default durable mid-scale
redis Higher claim throughput; set TRACE_QUEUE_URL
memory / nats (sim) Local tests only

Ingest and workers must share the same TRACE_QUEUE configuration. Soft backpressure rejects enqueue when ready depth exceeds the process limit (default 50 000).

Horizontal scaling checklist

  1. Run multiple workers against the same queue.
  2. Use S3-compatible object storage (not local disk) for multi-node.
  3. Put PgBouncer in front of Postgres for API fan-out.
  4. Scope tokens per project (X-Project-ID when multi-project).
  5. Tune per-project retention; GC runs in the worker process.
  6. Set TRACE_TRUSTED_PROXIES behind a real load balancer.

Not claimed

  • Sub-second analytics on trillions of events (no external warehouse required; optional NDJSON export exists).
  • Multi-region active-active without an external design.

Scrape GET /metrics with Prometheus; alert on job backlog growth.