generated: '2026-07-21'
method: searched
source: https://docs.superlog.sh
authentication:
  style: bearer
  key_types:
  - {name: management key, prefix: sl_management_, scope: org-wide, use: /api/v1/* REST}
  - {name: ingest key, prefix: sl_public_, legacy_prefix: superlog_live_, scope: single project, use: OTLP /v1/traces /v1/logs /v1/metrics}
  - {name: personal access token, prefix: superlog_pat_, scope: user + project, use: MCP server}
  ingest_header: 'x-api-key: <token> (or Authorization: Bearer <token>)'
  ref: authentication/superlog-authentication.yml
error_envelope:
  shape: '{ "error": "human-readable message" }'
  format: custom-non-rfc9457
  ref: errors/superlog-problem-types.yml
pagination:
  style: limit-only
  params: [limit]
  notes: >-
    Telemetry-read and MCP query tools cap rows with a `limit` parameter (e.g. 1-500 for
    logs/traces/metrics, 1-200 for incidents). No cursor/offset pagination documented.
request_idempotency:
  supported: false
  notes: >-
    The Management API documents no request Idempotency-Key header. Idempotency instead appears
    on the WEBHOOK RECEIVER side — each delivery carries a stable Superlog-Delivery UUID (reused
    across the 8-attempt retry schedule) intended as the consumer's de-duplication key.
webhooks:
  signing: HMAC-SHA256 over `<timestamp>.<rawBody>`, header Superlog-Signature (t=,v1=)
  replay_protection: reject timestamps drifting > 5 minutes
  delivery_id_header: Superlog-Delivery
  event_header: Superlog-Event
  retries: 8 attempts over ~8 hours (immediate, 30s, 1m, 2m, 5m, 15m, 1h, 6h)
  ref: asyncapi/superlog-webhooks.yml
versioning:
  ref: lifecycle/superlog-lifecycle.yml
content_type: application/json (management API); application/json or application/x-protobuf (OTLP ingest)