Skip to content

Overview

Kinglet collects logs from any app that can make an HTTP request, stores them in a single SQLite or libSQL database, and gives you a live-tailing UI with substring search — all on one small container.

Tiny by default

One process, one file, zero external services. ~150–250 MB of RAM, and millions of events a day fit comfortably on a 2 vCPU VPS.

Log from anything

NDJSON or JSON over HTTP. curl, Vector, Fluent Bit, the Node SDK — if it can POST, it can log to Kinglet.

Fast to search

SQLite FTS5 trigram indexing means substring matches ("timeout") just work, plus a live tail over server-sent events.

Alerts that don't cry wolf

Watch for instant matches, error-rate spikes, or a cron that went silent — delivered to a webhook, Discord, or Slack with cooldowns and auto-resolve.

Yours to run

Docker Compose in, Litestream backups out. MIT licensed, no phone-home, no telemetry.

Storage you choose

SQLite by default, libSQL for local files or hosted Turso, ClickHouse on the roadmap — one small interface, swappable drivers.

  • Node/PHP SDKs — a pino transport, a zero-code console preload, and a Monolog handler for Laravel/Symfony/WordPress.
  • More receivers — OTLP for OpenTelemetry pipelines, and syslog so nginx/php-fpm ship with zero code changes.
  • ClickHouse driver — the same storage interface, for when a single node stops being enough.