Skip to content
miniAPM

Getting Started

MiniAPM is the smallest useful APM—a self-hosted solution for monitoring application performance and tracking errors. Built in Rust with SQLite storage, it requires zero external dependencies.

Architecture

Section titled “Architecture”

MiniAPM consists of two services:

ServicePortDescription
miniapm3000Collector — Ingestion API for traces, errors, deploys
miniapm-admin3001Dashboard — Web UI for viewing and managing data

Both share the same SQLite database.

Quick Start

Section titled “Quick Start”

1. Run MiniAPM

Section titled “1. Run MiniAPM”
docker-compose.yml
services:
  miniapm:
    image: ghcr.io/miniapm/miniapm
    command: miniapm
    ports:
      - "3000:3000"
    volumes:
      - miniapm_data:/data
    environment:
      - SQLITE_PATH=/data/miniapm.db
      - RUST_LOG=mini_apm=info


  miniapm-admin:
    image: ghcr.io/miniapm/miniapm
    command: miniapm-admin
    ports:
      - "3001:3001"
    volumes:
      - miniapm_data:/data
    environment:
      - SQLITE_PATH=/data/miniapm.db
      - RUST_LOG=mini_apm_admin=info
      - ENABLE_USER_ACCOUNTS=true
      - SESSION_SECRET=change-me-to-random-string
    depends_on:
      - miniapm


volumes:
  miniapm_data:
Terminal window
docker compose up -d

On first run, check collector logs for your API key:

docker compose logs miniapm
INFO mini_apm::server: Single-project mode - API key: proj_abc123...

2. Configure Your Application

Section titled “2. Configure Your Application”

Point your OpenTelemetry SDK at the collector:

Terminal window
OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:3000/ingest
OTEL_EXPORTER_OTLP_HEADERS=Authorization=Bearer proj_abc123...

3. Open the Dashboard

Section titled “3. Open the Dashboard”

Navigate to http://localhost:3001 to see your traces and errors.

Core Concepts

Section titled “Core Concepts”

Traces & Spans

Section titled “Traces & Spans”

MiniAPM ingests traces via the OpenTelemetry Protocol (OTLP/HTTP JSON). Each request to your application creates a trace, which contains multiple spans representing individual operations (HTTP requests, database queries, external calls).

Error Tracking

Section titled “Error Tracking”

Errors are automatically captured with full stack traces and grouped by fingerprint to prevent duplicate noise. You can mark errors as resolved or ignored.

Route Performance

Section titled “Route Performance”

Every HTTP endpoint in your application gets performance metrics: P50, P95, P99 latencies, request counts, error rates, and database time breakdown.

What’s Next?

Section titled “What’s Next?”