Datadog

Datadog is a hosted observability platform combining metrics, logs, traces, RUM, synthetics, and cloud integrations. Agents run on hosts or as sidecars; libraries auto-instrument many runtimes.

| Area | Datadog piece | |------|----------------| | Metrics | Custom metrics, integrations | | APM | Distributed traces, service map | | Logs | Centralized ingestion and pipelines | | Monitors | Metric, log, trace, anomaly alerts |

Useful when teams want fast time-to-value without operating Prometheus/Grafana stacks—at vendor cost and agent overhead.

On interviews: agent vs serverless tracing, custom metric billing, and avoiding duplicate instrumentation with OTel.

Common pitfalls: custom metrics bill shock; logging everything at info; monitors without composite rollups.

The trade-off is managed velocity versus cost control and data egress policies.

Checklist:

• Scope custom metrics deliberately.
• Use pipelines to scrub PII in logs.
• Tag consistently for env/service.
• Review monitor noise monthly.