Databuff
GitHub
Skip to content

中文  |  English

Log Analytics · Design (v0.2.0+)

Version: 2026-07-03 · Status: design finalized
Backend: POST /log/search, Doris table log_dc_record (see OTLP Ingestion)

Background

v0.2.0 delivers OTLP Logs → Doris log_dc_record/log/search, with log tabs embedded in trace detail, service detail, and alert events.

A dedicated Log Analytics menu under Application Performance is planned as a global log search entry — APM-context exploration tied to traces and services, not a standalone ELK/Loki replacement.

Storage Model

Field areaPurpose
trace_id / span_idCorrelate with traces
service_id / hostnameFilter by service and host
severity / bodyLevel and message
attributes_jsonStructured OTel attributes

Ingest writes via OTLP /v1/logs (gRPC 4317 or HTTP 4318). See Telemetry Pipeline for the full pipeline.

API (MVP)

  • POST /log/search — keyword, service, host, trace/span, time range, offset pagination
  • POST /log/conditions — facet enumerations (evolving)

UI Direction

Reference patterns from SigNoz and Datadog: global time range, histogram, left facets (service / host / level), main log table with expand row and jump to trace when trace_id is present.

Out of scope for early releases: Live tail, LogQL/KQL-style query languages, log clustering.