Query and stream request logs
Understand the log schema, stream it into your own stack, and build SLO dashboards on top of it.
Last updated: September 11, 2026 · 7 min read
The log schema
Every log line is a JSON object describing exactly one request. The schema is versioned and strictly additive — fields are added over time, never removed or renamed.
json
{ "timestamp": "2026-09-24T04:31:07.412Z", "request_id": "req_01J8QK2M4X", "edge_region": "sin", "method": "GET", "path": "/v1/orders", "status": 200, "origin": "api-sin1.internal", "origin_latency_ms": 18.4, "cache_state": "hit", "retry_chain": []}Real-time streaming
Configure one or more sinks and logs start flowing within seconds. Failed deliveries are retried with exponential backoff for up to 24 hours.
Prometheus metrics
Scrape the metrics endpoint to build dashboards and alert on error budget burn. Every metric is labelled by zone, pool and edge region.
yaml
scrape_configs: - job_name: anylb metrics_path: /v1/metrics bearer_token: ${ANYLB_TOKEN} static_configs: - targets: ["metrics.anylb.com"]Saved queries
Save the filters your team reaches for during an incident, so everyone starts from the same view of the traffic.
Related reading