Rex - DataDog assistant

the SRE copilot for DataDog

Comment

Inspiration

It would be great if we could get notified and address incidents anytime, anywhere by having AI call you when it finds a critical incident, explain it to you in language and help you remediate by applying a playbook or giving you detailed instructions.

What it does

It’s like having an SRE sidekick in DataDog. Our assistant surfaces the right dashboards, metrics, logs, and alerts—so you can diagnose and resolve issues before they escalate, with fewer clicks and no noise.

How we built it

We integrated DataDog MCP server with VAPI to allow its assistant to provide DataDog information through web interfaces or phone to alert you and get incidents resolved or just get any DataDog monitoring information.

Challenges we ran into

  • Getting DataDog MCP server connected to VAPI with authentication.
  • Building a full end to end system including AWS, DataDog and VAPI for phone interactions.

🏆 Accomplishments That We’re Proud Of

  • Built a natural-language interface for observability: Users can ask Rex questions like “Why did latency spike yesterday?” or “What triggered this alert?”—and get actionable answers powered by DataDog data.

📚 What We Learned

  • Building workflows with DataDog MCP and other tools

🚀 What’s Next for Rex – DataDog Assistant

  • Support for applying remediation playbooks.
  • Deeper integrations with Slack, PagerDuty, and Jira: Let users escalate, share, or create tickets directly from Rex’s responses.
  • Role-based personalization: Tailor Rex’s responses for SREs, Security Analysts, or Engineering Managers with focused insights and dashboards.

Built With

Share this project:

Updates