Skip to content

Useful Docker Compose assistance

Some more ways LLMs can help with Docker issues

docker-compose files can be a lot of information.

Another way that I like to use LLMs to make working with Docker easier is to run prompts against docker-compose.yml files.

The methodology here is pretty self-explanatory:

  • Drop a docker compose into the LLM chat (or IDE copilot or whatever you're using)
  • Ask it questions about it

As I covered in a separate note, you can use LLMs to generate or modify existing docker compose files, like:

Could you add in this service? Could you change the database used from Postgres to Mongo (assuming it's supported)?

Example: which volumes have persistent data?

Something I like to make sure I'm always clear on when working with Docker is understanding which services have persistent mount points on the server and where those mount points are (ie, the paths).

Usually, I'll tweak the docker compose to set my own.

Providing an LLM with a docker compose I might prompt:

Read this docker compose. Identify which containers have persistent mount points. State where those are.

You can also specify the format you'd like to to output as, like:

Structure your ouput as follows:

Service (name of the service)

Local mount path