Datadog Agent Integrations
Overview
The Datadog Agent can collect host-level telemetry out of the box, but most of its value comes from integrations — also called checks.
An integration is a small, self-contained collector that knows how to gather telemetry from one specific technology: a database, a message broker, a web server, a cloud service, or an operating-system subsystem.
The Agent ships with a large catalog of checks — collectors for specific technologies — maintained in the Datadog integrations-core project.
Because Kloudfuse is protocol-compatible with Datadog, every Agent check works unchanged when you repoint the Agent at the Kloudfuse ingester: the telemetry a check produces flows into Kloudfuse alongside everything else the Agent gathers and becomes immediately queryable.
| Not every entry in the Datadog integrations tile is an Agent check. Integrations that deliver data through Datadog’s own backend — webhook integrations, API crawlers, and SaaS features such as LLM Observability (for example, incident.io, Contentful, and Anthropic) — are not covered here, because repointing the Agent does not redirect them. The Supported Integrations reference lists only Agent checks. |
How Integrations Work
Each integration runs as a scheduled check inside the Agent:
-
Discover the target — you tell the Agent where the technology is running, either with a static configuration file under
conf.d/<integration>.d/conf.yamlor, in Kubernetes, with Autodiscovery annotations on the pod so the Agent configures itself automatically as containers start and stop. -
Connect and collect — on each run (typically every 15 seconds) the check connects to the target — an admin endpoint, a status URL, a database query, a log file — and reads the current state.
-
Normalize and tag — the check converts what it read into Datadog-format metrics, logs, service checks, and events, and tags them with the integration’s
sourceso Kloudfuse applies the right processing pipeline. -
Forward to Kloudfuse — the Agent batches the results and sends them to the Kloudfuse ingester over HTTPS with the rest of its payload.
Supported Technologies
This is a categorized list of integrations by the kind of technology they monitor.
| Group | Examples |
|---|---|
PostgreSQL, MySQL, MongoDB, Elasticsearch, Oracle |
|
Redis, Memcached, Varnish |
|
Kafka, RabbitMQ, ActiveMQ, Pulsar |
|
Kubernetes, Docker, containerd, Argo, Nomad |
|
AWS, Azure, GCP managed services |
|
Cisco, Palo Alto, F5, SNMP devices |
|
Linux, Windows, systemd, host processes |
|
OpenAI, Anthropic, vLLM, Ray |
|
Cloudflare, Okta, CrowdStrike, audit sources |
|
Jenkins, GitLab, Terraform, Ansible |
|
PagerDuty, Slack, Statuspage |
|
Integrations that do not fall into the groups above |
You can find a comprehensive list of supported integrations in the reference pages.