Datadog
Also known as: DD, Datadog APM, Datadog agent
Datadog is a monitoring and security platform delivered as a SaaS product.
Install an agent on your servers or containers, and Datadog collects metrics, logs, and application traces, displaying them in unified dashboards. Unlike Prometheus+Grafana (self-managed), Datadog requires no infrastructure -- you pay per host and feature.
Datadog, explained at three levels
At home
Datadog is a paid service that watches your software and servers for you. Install a small agent program, and Datadog automatically starts tracking performance, errors, and logs -- visible in web dashboards accessible to your whole team.
At work
Companies use Datadog to monitor everything in their infrastructure from one place. Unlike building your own monitoring stack, Datadog is ready to use in hours with integrations for AWS, Kubernetes, databases, and major languages.
At home
Datadog's core products: Infrastructure Monitoring (host metrics), APM (application performance traces with flame graphs), Log Management (indexed logs with full-text search), Synthetics (simulated user journeys), RUM (Real User Monitoring for browser sessions), and Dashboards with auto-populated data from integrations.
At work
Datadog APM instruments code (via libraries for Python, Java, Node.js, Go, etc.) to capture distributed traces -- following a request from the frontend through microservices to the database. Service maps visualize dependencies. Monitors (alerts) fire when metrics deviate, with machine learning-based anomaly detection available.
At home
Datadog Logs uses a pipeline-and-processor model to parse, enrich, and route logs. Custom metrics via DogStatsD allow high-throughput metric submission. Datadog's proprietary Agent is open-source; it uses the OTLP protocol for OpenTelemetry compatibility alongside its own format.
At work
Datadog's pricing model (per host, per log volume, per APM host) makes cost management critical at scale. Teams use Metrics without Limits to control cardinality costs and log archives (cheap object storage) with rehydration for compliance without paying for live indexing of all logs. Datadog Bits AI adds natural language querying of dashboards and logs.
Real-world examples
At home
Jerome's company installs the Datadog agent on all their servers. When their API response time spiked last Tuesday, the APM flame graph showed exactly which database query was the bottleneck -- found and fixed in 20 minutes.
At work
Nadia's platform team uses Datadog as their single pane of glass: AWS metrics from CloudWatch integration, container metrics from the Kubernetes agent, application traces from APM, and security signals from Datadog SIEM -- all in one dashboard, one alert system, one bill.
Common confusions
- Datadog is just a metrics dashboard like Grafana — Datadog is a full observability platform: metrics, logs, traces (APM), security monitoring, synthetic testing, and incident management. Grafana is primarily a visualization layer for data you manage separately.
- Datadog is always expensive — Datadog's pricing scales with usage. Small teams with a few hosts pay modest amounts. It gets expensive at scale -- which is why large organizations carefully manage which hosts and features they instrument.
For analysts
Business analysts can use Datadog dashboards to track product metrics -- API usage by customer, error rates by feature, and response time trends -- using the same tool the engineering team uses, with read-only access for non-engineers.
History & origin
Datadog was founded in 2010 by Olivier Pomel and Alexis Le-Quoc, who met while working at a startup struggling with infrastructure visibility. The company went public in 2019. Datadog pioneered unified observability -- combining metrics, logs, and distributed traces (APM) in one product when competitors offered them separately. It expanded into network monitoring, security (SIEM, CSPM), synthetics, and CI visibility, becoming a broad platform for engineering operations.