Why are agents installed at runtime? #16132
-
Trying to optimise load times a bit and was trying to determine why the agent is set up at runtime, rather than being preinstalled into a docker image? I'm assuming it's due to not wanting to maintain a build pipeline for docker image files, but I wanted to double check. If I'm always deploying to a fixed target, would there be any issue with pre building an agent into a dockerfile? The agent isn't compiled by the host system and baking some unique keys or something into the runtime that's used to communicate between the running services? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
You are correct. This also helps ensure that agent and server versions are the same. Now, answering your question, yes, you can surely back the agent into the Docker image and run that as a startup and make sure you update the |
Beta Was this translation helpful? Give feedback.
You are correct. This also helps ensure that agent and server versions are the same.
Now, answering your question, yes, you can surely back the agent into the Docker image and run that as a startup and make sure you update the
CODER_AGENT_TOKEN
as it refreshed on each start.