Aside from the projects mentioned in this open source hub, the Datadog related projects that collect telemetry data from Datadog’s customers’ infrastructure are also open source. We will explore some of those in this spotlight article.
The Datadog Agent is a Go application that is deployed in every host that needs to be monitored in Datadog. The Datadog Agent is distributed as several different Linux distributions packages, as Windows and MacOS packages, and as Docker containers, ensuring that all types of infrastructure can be properly monitored.
The Datadog Agent is composed of several agents, including the main agent, the tracing agent, the process agent, and the security agent.
One of the key features of Datadog is the +600 integrations that help Datadog’s customers get visibility into their systems in a very simple way. These built-in turn-key integrations include Agent integrations (or checks). The Datadog Agent schedules and runs these checks in Datadog’s customers’ infrastructure.
All Agent integrations are open source and there are two main repositories with all their code:
These integrations are maintained by Datadog and they are already built-in with the Datadog Agent distribution packages, so customers can start using those integrations as soon as they deploy the Agent.
These integrations are community-driven integrations that are not maintained by Datadog, but that use the same CI/CD infrastructure as integrations-core to verify and test them. These are not shipped with the Agent and need to be explicitly installed by Datadog’s customers.
Coding, maintaining, and testing all these integrations is not an easy task. Julia Simon and Christine Chen explained how their team does this in a Datadog on episode.
Tracers are language specific libraries that allow developers to instrument their code to collect traces that will be sent to Datadog by the Agent.
Interact and integrate with mobile, browser, and other platforms with our open source SDKs.