Agent Skills: Sentry Deploy Integration

Sentry Deploy Integration

Overview

Wire Sentry into your deploy pipeline so every release is tracked end-to-end: commit association, source map upload, deploy recording, and post-deploy health monitoring. Sentry links errors to the exact deploy and suspect commit that introduced them, giving you crash-free session rates, adoption curves, and regression alerts per release.

Prerequisites

• Sentry CLI installed (npm i -g @sentry/cli or curl -sL https://sentry.io/get-cli/ | bash)
• SENTRY_AUTH_TOKEN with project:releases and org:read scopes
• SENTRY_ORG and SENTRY_PROJECT environment variables set
• @sentry/node v8+ installed in your application
• Source control integration enabled in Sentry (Settings > Integrations > GitHub/GitLab)

Instructions

Step 1 --- Record Deploys with sentry-cli

Create a release, associate commits for suspect-commit linking, and record the deployment with timing metadata.

#!/bin/bash
# scripts/sentry-deploy.sh
set -euo pipefail

VERSION="${1:-$(sentry-cli releases propose-version)}"
ENVIRONMENT="${2:-production}"
DEPLOY_START=$(date +%s)

# Create release and associate commits (enables suspect commits)
sentry-cli releases new "$VERSION"
sentry-cli releases set-commits "$VERSION" --auto

# Upload source maps for readable stack traces
sentry-cli sourcemaps upload \
  --release="$VERSION" \
  --url-prefix="~/static/js" \
  --validate \
  ./dist

# Finalize marks the release as ready
sentry-cli releases finalize "$VERSION"

# --- Deploy your application here ---
# e.g., kubectl set image deployment/app app=myapp:$VERSION

DEPLOY_END=$(date +%s)

# Record the deployment in Sentry with timestamps
sentry-cli releases deploys "$VERSION" new \
  -e "$ENVIRONMENT" \
  -t "$DEPLOY_START" \
  -f "$DEPLOY_END"

echo "Sentry deploy recorded: $VERSION -> $ENVIRONMENT ($(( DEPLOY_END - DEPLOY_START ))s)"

For multi-environment promotion (staging then production):

# Stage 1: deploy to staging
sentry-cli releases deploys "$VERSION" new -e staging

# Stage 2: after QA passes, deploy to production
sentry-cli releases deploys "$VERSION" new -e production

# Sentry dashboard shows the full promotion timeline

Step 2 --- Tag Releases in the SDK and Monitor Health

Configure the Sentry SDK with the release tag so crash-free session/user metrics, adoption rates, and error attribution bind to each release.

// src/instrument.ts
import * as Sentry from '@sentry/node';

Sentry.init({
  dsn: process.env.SENTRY_DSN,
  release: process.env.SENTRY_RELEASE,   // e.g. "my-app@1.4.2"
  environment: process.env.NODE_ENV,      // "production" | "staging"

  // Release health: session tracking is on by default in v8
  // Crash-free sessions/users calculated automatically

  // Sample 10% of transactions in production
  tracesSampleRate: process.env.NODE_ENV === 'production' ? 0.1 : 1.0,
});

Release health metrics Sentry tracks automatically:

| Metric | What it measures | |--------|-----------------| | Crash-free sessions | % of sessions with zero unhandled errors | | Crash-free users | % of distinct users with zero unhandled errors | | Adoption | % of sessions on this release vs. the previous release | | Error count | New and total errors attributed to this release | | Session count | Total sessions observed on this release |

Step 3 --- Compare Releases, Detect Regressions, and Configure Notifications

Compare releases via the API to catch regressions before they spread:

# List recent releases with new-issue counts and deploy info
curl -s \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  "https://sentry.io/api/0/organizations/$SENTRY_ORG/releases/?project=$SENTRY_PROJECT&per_page=5&health=1" \
  | python3 -c "
import json, sys
releases = json.load(sys.stdin)
for r in releases:
    health = r.get('healthData', {})
    crash_free = health.get('crashFreeSessions', 'N/A')
    new_issues = r.get('newGroups', 0)
    deploys = len(r.get('deploys', []))
    print(f\"{r['version']}: {new_issues} new issues, {deploys} deploys, crash-free: {crash_free}\")
"

Suspect commits --- when set-commits is configured, Sentry traces each new error to the commit that likely introduced it:

# Fetch suspect commits for a specific issue
curl -s \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  "https://sentry.io/api/0/issues/$ISSUE_ID/committers/" \
  | python3 -c "
import json, sys
data = json.load(sys.stdin)
for c in data.get('committers', []):
    author = c['author']['name']
    for commit in c['commits']:
        print(f\"  Suspect: {commit['id'][:8]} by {author} — {commit['message'].splitlines()[0]}\")
"

Deploy notification webhooks --- configure in Project Settings > Integrations > Internal Integrations or use the webhook directly:

// routes/sentry-webhook.ts
import express from 'express';
const router = express.Router();

interface SentryDeployPayload {
  action: string;
  data: {
    deploy: {
      environment: string;
      dateFinished: string;
    };
    release: {
      version: string;
      projects: Array<{ name: string }>;
    };
  };
}

router.post('/webhook/sentry-deploy', (req, res) => {
  const payload = req.body as SentryDeployPayload;

  if (payload.action === 'deploy.created') {
    const { release, deploy } = payload.data;
    console.log(`Deploy: ${release.version} -> ${deploy.environment}`);
    // Trigger post-deploy smoke tests or Slack notification
  }

  res.status(200).json({ ok: true });
});

export default router;

Rollback tracking --- record the rollback as a new deploy pointing to the previous stable version:

ROLLBACK_TO="my-app@1.3.9"
sentry-cli releases deploys "$ROLLBACK_TO" new \
  -e production \
  --name "Rollback from $CURRENT_VERSION"
# Sentry attributes new errors to the rolled-back version

Output

After completing these steps you will have:

• Every deploy recorded in Sentry with environment and timestamps
• Crash-free session and user rates tracked per release
• Adoption curves showing rollout progress
• Suspect commits linking new errors to the exact commit that introduced them
• Release comparison showing regression counts across versions
• Deploy notifications flowing to Slack, PagerDuty, or custom webhooks
• Rollback events visible in the release timeline

Error Handling

| Error | Cause | Solution | |-------|-------|----------| | error: release not found | Deploy created before release exists | Run sentry-cli releases new $VERSION before recording the deploy | | No release health data | Session tracking disabled or SDK < v7.2 | Upgrade to @sentry/node v8+; do not set autoSessionTracking: false | | Wrong environment on events | environment not set in SDK init | Pass environment explicitly in Sentry.init() | | Suspect commits missing | Source control integration not linked | Enable GitHub/GitLab in Settings > Integrations and run set-commits --auto | | 401 Unauthorized on deploy API | Token missing project:releases scope | Regenerate token at https://sentry.io/settings/account/api/auth-tokens/ | | Crash-free rate stuck at 100% | Release tag mismatch between CLI and SDK | Ensure SENTRY_RELEASE in Sentry.init() matches the sentry-cli releases new version exactly | | Deploy timestamps zero | Missing -t/-f flags on deploys new | Capture $(date +%s) before and after deploy, pass both flags |

Examples

TypeScript: Full deploy script with health check polling

// scripts/deploy-and-monitor.ts
import { execSync } from 'child_process';

const VERSION = process.env.SENTRY_RELEASE || execSync('sentry-cli releases propose-version').toString().trim();
const ENV = process.env.DEPLOY_ENV || 'production';
const ORG = process.env.SENTRY_ORG!;
const TOKEN = process.env.SENTRY_AUTH_TOKEN!;

async function checkReleaseHealth(version: string): Promise<void> {
  const res = await fetch(
    `https://sentry.io/api/0/organizations/${ORG}/releases/${encodeURIComponent(version)}/`,
    { headers: { Authorization: `Bearer ${TOKEN}` } }
  );
  const release = await res.json();
  const crashFree = release.healthData?.crashFreeSessions;
  console.log(`Release ${version} crash-free sessions: ${crashFree ?? 'pending'}%`);

  if (crashFree !== undefined && crashFree < 95) {
    console.error(`ALERT: Crash-free rate ${crashFree}% is below 95% threshold`);
    process.exit(1);
  }
}

// Record deploy
execSync(`sentry-cli releases deploys "${VERSION}" new -e ${ENV}`, { stdio: 'inherit' });

// Poll health after deploy
setTimeout(() => checkReleaseHealth(VERSION), 5 * 60 * 1000);

CLI: GitHub Actions integration

# .github/workflows/deploy.yml
name: Deploy with Sentry
on:
  push:
    branches: [main]

jobs:
  deploy:
    runs-on: ubuntu-latest
    env:
      SENTRY_AUTH_TOKEN: ${{ secrets.SENTRY_AUTH_TOKEN }}
      SENTRY_ORG: my-org
      SENTRY_PROJECT: my-app
    steps:
      - uses: actions/checkout@v4
        with: { fetch-depth: 0 }  # Full history for set-commits

      - run: npm ci && npm run build

      - name: Create Sentry release
        run: |
          VERSION="my-app@${{ github.sha }}"
          npx @sentry/cli releases new "$VERSION"
          npx @sentry/cli releases set-commits "$VERSION" --auto
          npx @sentry/cli sourcemaps upload --release="$VERSION" ./dist
          npx @sentry/cli releases finalize "$VERSION"

      - name: Deploy to production
        run: ./scripts/deploy.sh

      - name: Record deploy in Sentry
        run: |
          VERSION="my-app@${{ github.sha }}"
          npx @sentry/cli releases deploys "$VERSION" new -e production

Resources

• Sentry Release Setup
• Release Health Metrics
• sentry-cli Deploys
• Releases API
• Suspect Commits
• Deploy Notifications

Next Steps

For error monitoring configuration, see sentry-error-monitoring. For performance tracing setup, see sentry-performance-tracing.