Open source alternatives to Datadog

What Datadog does

Datadog is a software-as-a-service platform that collects telemetry through its own agent, more than a thousand integrations and OpenTelemetry. Its catalog spans infrastructure and container monitoring, custom metrics, APM and distributed tracing, continuous profiling, database monitoring, log management, error tracking, real user monitoring and session replay, synthetic tests, security products, incident management and AI assistants. There is no self-hosted edition: data is stored on one of Datadog’s sites in the US, the EU (Germany), Japan, Australia or the UK.

Why people look for an alternative

• Many separately metered products. Each product has its own billing unit: hosts per month for infrastructure and APM, millions of indexed log events, gigabytes ingested and others. Datadog counts hosts every hour and bills on the high-water mark of the lower 99 percent of those hours, so short-lived hosts and autoscaling show up on the bill. The total depends on many separate usage figures.
• Limited free plan. The free infrastructure plan covers up to five hosts with one day of metric retention; the paid plans keep metrics for 15 months.
• Retention set per product. Defaults include 15 days for APM errors and 30 days for browser RUM sessions, and longer log retention is a separately priced tier.
• Data location. Each Datadog site is independent, and data cannot be shared across sites. Your telemetry, which can include log lines and request details, stays on the vendor’s infrastructure.
• Lock-in. Dashboards, monitors and data collected with the proprietary agent do not move easily to another tool.

What to compare in an alternative

1. Coverage. Which Datadog products you actually use: APM and traces, logs, infrastructure metrics, RUM, synthetic checks, error tracking or incident management.
2. Instrumentation. OpenTelemetry support, and whether there is a bridge that accepts traffic from existing Datadog agents during a migration.
3. Storage and scaling. ClickHouse, Parquet on object storage or Prometheus, and the hardware each needs at your volume.
4. Retention control. Self-hosting means retention is limited by your storage budget rather than by a plan.
5. What is in the open source edition. SSO, RBAC and audit logs are often in paid editions.

Which one to choose

SigNoz is the closest match for APM: traces, logs and metrics in one interface on ClickHouse, built on OpenTelemetry, with migration guides from Datadog, a collector receiver for Datadog agent traffic and a dashboard translation tool. Its enterprise directories are proprietary, and SSO is a paid feature. OpenObserve is strongest on logs at volume, storing logs, metrics, traces and RUM data as Parquet files on local disk or object storage and querying them with SQL or PromQL; it runs as a single binary for small setups. Coroot collects metrics, logs, traces and profiles with an eBPF agent and no code changes, which suits Kubernetes teams that want a service map and guided root cause checks. OneUptime combines OpenTelemetry-based observability with uptime checks, status pages, incidents and on-call scheduling, so it can also replace tools that sit next to Datadog; it is the heaviest of the four to self-host.

If you only need infrastructure metrics and alerts, a stack of Prometheus and Grafana, or Zabbix or Netdata, may be enough. See the monitoring guide for how these tools differ.