FAQs related to Docker application and Docker agent
What is the necessity to have an exclusive docker agent?
When changing from monolithic application deployments to cloud native microservices using AWS ECS, Kubernetes etc., an exclusive monitoring agent would prove to be more secure and flexible to keep track of your docker containers and hosts. Also, the agent follows the "container philosophy" i.e., any single active process is wrapped inside a container, thus making cluster management more easy.
How does installation process vary between the docker agent and the docker application? The docker application will be auto discovered once the Linux agent is installed. On the other hand, the docker agent has to be installed in the respective container to monitor the underlying host.
When installed as an agent, will the server also be monitored? Yes, once the docker agent is installed, a monitor will be created for the server and another monitor would be created for the docker.
Can the docker agent be installed in a machine, where there is already a Linux agent? No, either the Linux or the Docker agent can run in a machine.
Will there be high memory and CPU usage because of the docker agent? The docker agent needs a minimum of 256 Mi of CPU and 200 MB of memory. The user can also set limits as to how much memory and CPU should be consumed by the agent.
Will there be a change in the performance metrics while using the docker application and the docker agent? No, the performance metrics will remain the same, except for disk related metrics. To get disk related metrics for the docker agent, mount the volumes into the container and get them monitored.
Is there any licensing difference between the docker agent and the docker application? No, there is no licensing difference. Irrespective of whether an application or agent, docker is a basic monitor.
Any bulk installation techniques for the docker agent?