G1Joshi Datadog
Datadog is a leading SaaS observability platform. In 2025, it focuses on AI Observability (monitoring LLMs) and automated remediation with Watchdog.
When to Use
- SaaS Convenience: You want a complete solution (APM, Logs, Infra) without managing storage.
- Full Stack Visibility: Frontend RUM (Real User Monitoring) connected to Backend Traces connected to DB Metrics.
- AI Apps: Monitor token usage, latency, and costs of LLM calls.
Quick Start
Install Agent:
DD_API_KEY=... bash -c "$(curl -L https://s3.amazonaws.com/dd-agent/scripts/install_script.sh)"
Enable APM (e.g. Node.js):
DD_TRACE_AGENT_URL=http://localhost:8126 node --require dd-trace/init app.js
Core Concepts
Tags
The most important concept. env:prod, service:login, team:core. Filter everything by tags.
Watchdog
AI-driven anomaly detection. "Redis latency is 30% higher than normal".
APM (Application Performance Monitoring)
Automatic instrumentation of code to find slow SQL queries or API calls.
Best Practices (2025)
Do:
- Tag Everything: Use
DD_TAGSto standard metadata across all hosts. - Use Sampling: For high-volume services, sample traces to keep costs down.
- Set Budgets: Datadog is expensive. Use cost alerts.
Don't:
- Don't ignore the bill: Custom Metrics and high-volume logs can spike costs unexpectedly.