Sentry — Open Source Error Tracking & Performance Monitoring

What Sentry Does

• Error Tracking: Capture and group unhandled exceptions with full stack traces
• Performance Monitoring: Trace transactions across services with latency breakdowns
• Session Replay: Replay user sessions to see exactly what happened before an error
• Release Tracking: Track error rates per release, detect regressions automatically
• Source Maps: Unminify JavaScript errors with automatic source map processing
• Alerts: Configurable alert rules for error spikes, new issues, and performance degradation
• Issue Assignment: Auto-assign issues to code owners based on stack trace
• Integrations: GitHub, GitLab, Jira, Slack, PagerDuty, and 50+ integrations
• Breadcrumbs: Automatic event trail leading up to each error (clicks, API calls, console logs)

SDK Examples

JavaScript/Node.js

const Sentry = require("@sentry/node");
const { nodeProfilingIntegration } = require("@sentry/profiling-node");

Sentry.init({
  dsn: "https://your-dsn@sentry.io/project",
  integrations: [nodeProfilingIntegration()],
  tracesSampleRate: 0.1,
  profilesSampleRate: 0.1,
});

// Errors are captured automatically
// Manual capture:
try {
  riskyOperation();
} catch (error) {
  Sentry.captureException(error);
}

// Add context
Sentry.setUser({ id: "user123", email: "user@example.com" });
Sentry.setTag("feature", "checkout");

Python

import sentry_sdk

sentry_sdk.init(
    dsn="https://your-dsn@sentry.io/project",
    traces_sample_rate=0.1,
)

# Auto-captures unhandled exceptions
# Django, Flask, FastAPI auto-instrumented

# Manual capture
try:
    process_payment(order)
except PaymentError as e:
    sentry_sdk.capture_exception(e)
    sentry_sdk.capture_message("Payment failed for order %s" % order.id)

Go

import "github.com/getsentry/sentry-go"

func main() {
    sentry.Init(sentry.ClientOptions{
        Dsn: "https://your-dsn@sentry.io/project",
        TracesSampleRate: 0.1,
    })
    defer sentry.Flush(2 * time.Second)
}

Self-Hosting

Requirements

• Docker 19.03+ and Docker Compose 2.19+
• 8GB RAM minimum (16GB recommended)
• 20GB disk space

Docker Compose

git clone https://github.com/getsentry/self-hosted.git
cd self-hosted
./install.sh  # Interactive setup
docker compose up -d

Components deployed:

• Sentry Web (Django)
• Sentry Worker (Celery)
• Sentry Cron
• PostgreSQL
• Redis
• Kafka
• ClickHouse
• Snuba (query engine)
• Symbolicator (source maps)
• Relay (event ingestion)

Key Features

Issue Grouping

Sentry automatically groups similar errors:

Issue: TypeError: Cannot read property 'name' of undefined
├── 1,234 events
├── 56 users affected
├── First seen: 2 days ago
├── Last seen: 5 minutes ago
├── Stack trace with source context
├── Breadcrumbs (30 events before crash)
├── Tags: browser=Chrome, os=Windows, release=v2.1.0
└── Linked commits (suspect commits)

Performance Tracing

Transaction: POST /api/checkout (p50: 450ms, p95: 1.2s)
├── HTTP Handler (12ms)
├── Auth Middleware (5ms)
├── Validate Cart (8ms)
├── DB: SELECT products (45ms)
├── Payment API Call (380ms) ← bottleneck
├── DB: INSERT order (15ms)
├── Send Email (async)
└── Response (465ms total)

Release Health

Release v2.1.0:
├── Crash-free sessions: 99.2% (↓ from 99.8%)
├── Crash-free users: 98.5%
├── New issues: 3
├── Regressed issues: 1
├── Adoption: 45% of users
└── Suspect commits: abc1234 by @developer

Sentry vs Alternatives

FAQ

Q: Is self-hosted Sentry resource-intensive? A: Yes — a full Sentry deployment includes Kafka, ClickHouse, PostgreSQL, and other components, requiring at least 8 GB RAM. For small teams, consider the lightweight alternative GlitchTip (Sentry-compatible but with much lower resource needs).

Q: Does the self-hosted version have the same features as SaaS? A: Core functionality (error tracking, performance monitoring, alerting) is identical. The SaaS edition adds advanced features like Session Replay and Cron Monitoring, plus better scalability and zero maintenance.

Q: Does performance degrade at high data volumes? A: The self-hosted version should use event sampling (tracesSampleRate) to control data volume. For high-traffic apps, sampling rates of 0.01–0.1 give enough insight. The ClickHouse backend keeps query performance strong.

Sources & Credits

• GitHub: getsentry/sentry — 43.5K+ ⭐
• Self-Hosted: getsentry/self-hosted
• Website: sentry.io