A lightweight, self-hosted service for monitoring application errors and logs.
It collects exceptions and logs from your backend and frontend apps through the official Sentry SDKs: you point an SDK at it with a DSN, no other client library is needed. Repeated errors are grouped into issues, your team is notified when something breaks, and coding agents can read the errors over MCP to help explain them. Everything runs in one small container on your own server, so your data never leaves your infrastructure.
- Errors and logs from any app with a Sentry SDK (Node.js, browsers, Python, …), several projects per instance.
- Issues: repeated errors are grouped, with resolve/ignore, regression detection and a history of who changed each status.
- Logs with search, linked to the error of the same request.
- Notifications about new and regressed issues by email, Telegram and Slack, with retries that survive restarts.
- Web UI with users and roles (superadmin, admin, user).
- MCP server for coding agents (Claude Code, Codex, …) with OAuth sign-in or a personal token.
- REST API for scripts and integrations.
- Retention limits by age, count and database size; credentials are scrubbed from incoming data.
- Small: one static binary with embedded SQLite, an image of about 20 MB, no external database.
Tracing spans, session replays and other Sentry item types are accepted and dropped.
docker run -d --name exception-monitoring -p 8000:8000 -v exception-monitoring-data:/data \
-e ADMIN_PASSWORD='<at least 12 characters>' \
devforth/exception-monitoring:latestOpen http://localhost:8000 and log in as admin with that password. On the Projects page add a project and copy its DSN.
For a real deployment use compose.yml with an .env file and put the server behind HTTPS: see Deployment.
Install the Sentry SDK of your platform and give it the project's DSN, e.g. for Node.js:
import * as Sentry from "@sentry/node";
Sentry.init({
dsn: "https://<public_key>@errors.example.com/1",
enableLogs: true,
});More platforms, browser apps and what is stored: Connecting apps.
skills/exception-monitoring teaches a coding agent to deploy the server, connect itself over MCP, add the Sentry SDK to your apps and investigate errors, following these docs. Install it once for your user:
npx skills add devforth/exception-monitoring -gOr copy the folder to ~/.claude/skills/ (Claude Code) or ~/.agents/skills/ (Codex). Then ask, e.g., "connect this app to our error monitor".
- Deployment: Docker, reverse proxy and HTTPS, upgrades.
- Configuration: every environment variable.
- Settings in the web UI: the fields of the Projects and Notifications pages.
- Connecting apps: SDK setup, levels, logs, grouping.
- Coding agents (MCP): connecting Claude Code, Codex and other clients.
- REST API.
- Sentry protocol notes: the envelope format as SDKs send it.
Needs Go (see go.mod) and golangci-lint v2.
cp .env.example .env.local # then set ADMIN_PASSWORD and the other values
make run # http://localhost:8000
make test lint # what CI runs
make docker image-size # build the image and check its sizePushing a tag vX.Y.Z publishes the image devforth/exception-monitoring for amd64 and arm64.
MIT © Devforth.io