Logflare

Ingest API: POST /logs with source parameter and log event body. Log events are JSON objects with optional metadata. Sources organize log streams — each source maps to a BigQuery table. The ingest endpoint is intentionally simple: one endpoint, one purpose. Query API: BigQuery SQL via the BigQuery API or Logflare's query endpoint. Rules define log routing conditions (match events to destinations). The API surface is minimal: create sources, ingest logs, query via SQL. No built-in alerting, dashboards, or APL-style query language — these capabilities come from BigQuery or connected tools. For agents wanting simple structured log storage with SQL querying, the API is appropriately minimal.

API errors return JSON with message. Standard HTTP status codes. Ingest errors include validation messages for malformed events. Rate limits depend on the plan and BigQuery quotas. The main reliability consideration: Logflare depends on BigQuery — BigQuery availability and quotas affect log querying. Ingest buffering handles temporary BigQuery unavailability. Query performance depends on BigQuery's query execution engine — simple queries are fast, complex aggregations over large time ranges may be slower. For agents, the BigQuery dependency is both a feature (SQL power, scalable storage) and a constraint (BigQuery pricing, quota management). Self-hosted Logflare requires a BigQuery project.

Logflare provides log management backed by BigQuery — logs are ingested via a simple API and stored in BigQuery tables for SQL-based querying. Acquired by Supabase, it now powers Supabase's built-in logging infrastructure. For agents, the value proposition is straightforward: send structured log events via HTTP, query them with SQL. The simplicity compared to Axiom, Datadog, or Elasticsearch is intentional — Logflare covers the core log-ingest-query workflow without the complexity of a full observability platform. Sources organize log streams (one source per application/service). Rules can route logs to different sinks. For agents in the Supabase ecosystem, Logflare is the native logging solution. For agents outside Supabase, the BigQuery dependency and simpler feature set may be limitations.

API key authentication via X-API-KEY header or api_key parameter. Source tokens provide per-source ingest access. The source token model means agents can have ingest credentials for specific log streams without management access. No OAuth. Keys don't expire. For Supabase users, Logflare authentication is integrated with the Supabase project. The auth model is simple — appropriate for the tool's scope. Management operations (source creation, rule configuration) use the account-level API key.

Documentation at logflare.app/guides covers setup, API reference, and integration guides. The documentation is adequate but less comprehensive than Axiom or Datadog — reflecting the tool's simpler scope. Supabase integration documentation explains how Logflare powers Supabase's logging features. Client libraries: Elixir (official), JavaScript/TypeScript, and community libraries for other languages. The API's simplicity means documentation needs are modest. For agents in the Supabase ecosystem, the Supabase logging documentation is more relevant than Logflare's standalone docs. For agents using Logflare independently, the API reference and source configuration guides are the starting points.