Best practices
As you have seen, there are multiple ways to install and configure the Datadog Agent, and, for someone new to Datadog, it could be daunting to determine how the agent can be rolled out and fine-tuned efficiently to meet the monitoring requirements. However, there are a few things that are obvious as best practices, and let's summarize those here:
- If the agent is installed on the host, plan to include it in the machine image used to spin up or boot the host.
- Set up Ansible playbooks or similar tools to make ad hoc changes to the Datadog Agent on the host. This is not recommended for some complex infrastructure environments, especially where bare-metal servers are used, so some in-place change might be needed.
- When containers are to be monitored, plan to deploy the agent also as a container.
- Plan to collect tags from underlying infrastructure components such as Docker and Kubernetes by suitably configuring the agent.
