Newshot

Replit Datacat – their self-hosted observability is AI slop and it doesn't work

Original text(via HN AI)

DataCat

Self-hosted observability: metrics, logs, distributed traces, dashboards, infrastructure and alerts in one place. DataCat runs on your infrastructure, stores data in PostgreSQL (or telemetry in ClickHouse, if you choose; see docs/SELF_HOSTING.md), and accepts OpenTelemetry and Datadog-compatible agent traffic. No hosted account or external service is required.

Features

  • Metrics explorer, formulas, dashboards and host maps
  • Searchable logs with facets, live tail and trace correlation
  • APM service catalog, trace explorer and waterfall
  • Metric, log and trace monitors with webhook/Slack notifications
  • Organizations, local accounts, single-use invites, API/application keys
  • OpenTelemetry Collector-based host agent and Node.js/Python instrumentation libraries
  • One production process serving the API, web UI, installer and downloads
DataCat dashboard
DataCat dashboard

Quick start with Docker Compose

Requirements: Docker Engine with the Compose plugin, Git, and OpenSSL.


        git clone https://github.com/replit/datacat
cd datacat
cp .env.example .env
sed -i "s/^POSTGRES_PASSWORD=.*/POSTGRES_PASSWORD=$(openssl rand -hex 24)/" .env
sed -i "s/^SESSION_SECRET=.*/SESSION_SECRET=$(openssl rand -hex 32)/" .env
chmod 600 .env
mkdir -p downloads
docker compose up -d --build
      

Open http://localhost:3000, create your admin account and organization, then create an API key under Organization Settings. Setup asks for a setup token, printed in the server log (docker compose logs datacat | grep "setup token") unless you set DATACAT_SETUP_TOKEN. Set PUBLIC_URL to your externally reachable HTTPS origin before inviting colleagues or installing agents remotely.

Agent binaries and SDK packages are separate release assets. Put them (including SHA256SUMS) in downloads/ to serve them from your instance.

Sending data

DataCat accepts the Datadog public intake API and OTLP/HTTP:

SenderStatus
Datadog Agent 7.56.2 (metrics, logs, check runs, traces)Tested in CI-style runs
Datadog Agent 7.85.4 (metrics, logs, check runs, traces)Tested with captured payloads; including the newer string-table trace format
Datadog Agent 7.57–7.84Not tested individually. Agents that send traces in a format DataCat cannot read get a 422, never a silent 202
OpenTelemetry Collector / SDKs (OTLP/HTTP)/v1/traces , /v1/metrics , /v1/logs (also under /otlp/v1/* )
DataCat Agent and librariesTested in CI

GET /api/v1/intake/stats shows, per intake path, how many items were stored and how many were dropped since the server started.

Sign-in uses local email/password accounts; members join through single-use invite links.

From source

Use Node.js 22.15+, pnpm 10.11.0 and PostgreSQL 14+. Node 22 also meets the Node SDK requirement; older Node 20 releases are not supported by the current SDK.


        corepack enable
corepack prepare [email protected] --activate
pnpm install --frozen-lockfile
BASE_PATH=/ pnpm run build:selfhost
# Set DATABASE_URL, SESSION_SECRET and PUBLIC_URL in your environment.
NODE_ENV=production PORT=3000 pnpm start
      

The server migrates its schema automatically before listening. There is no manual schema push and no separate frontend server.

See Self-hosting for clean Ubuntu installation, all environment settings, backups, systemd, agent installation, Node/Python instrumentation and a webhook monitor walkthrough.

Development


        pnpm run typecheck:libs
pnpm --filter @workspace/api-server run dev
# In another terminal:
API_PROXY=http://localhost:3000 PORT=5173 pnpm --filter @workspace/web run dev
      

The API needs DATABASE_URL; use a separate development database. The agent is built with Go 1.26.11+:


        bash agent/build.sh
DESTDIR="$PWD/dist/downloads" bash agent/release.sh
      

        pnpm --filter @workspace/api-server test     # server unit tests
DATACAT_URL=http://localhost:3000 DATACAT_SETUP_TOKEN=<token> bash scripts/e2e.sh   # end-to-end known-answer checks against a fresh install
      

CI (.github/workflows/ci.yml) runs typechecks, unit tests, dependency and vulnerability audits, the end-to-end checks on both storage options, the library and agent tests, a Docker build and a secrets scan on every pull request.

Runnable examples are in examples/. See Contributing, Security and the Changelog.

Benchmarks

bench/kit sends metrics, logs and traces through DataCat and checks every answer against a known value; it reports throughput, query latency and how long data takes to become searchable, for both storage options. The other folders in bench/ hold the correctness, resilience, alerting, soak and isolation tests; raw results behind the published numbers are in bench/evidence/.

License

Apache License 2.0. Third-party dependencies retain their own licenses. DataCat is an independent project, not affiliated with or endorsed by Datadog, Inc. Datadog is a trademark of Datadog, Inc.; it is named here only to describe protocol compatibility.

The text above was extracted automatically from the source page and may miss images, tables or code blocks. The original is authoritative.

Published by HN AI
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗