docker-compose-optimization

[vanodevium]

Hi! This project looks awesome, really!

I try to optimize docker-compose.yml for common use cases:

• there is no need for container names because internal docker dns uses service names as endpoint also
• also there is no need to change internal service ports, please support changes only for exposed ports, not internal
• docker-compose version tag is deprecated now
• set project name for global namespace
• there is 2025 year... .env file have to be in .gitignore

I also think that there is no point in cloning the repository to run this project.

So it is enough to describe only two points in the readme:

• how to use curl and download docker-compose.yml only
• what variables can be set into the .env file

[adanb13]

Hi, just wanted to note that we have seen this and will get to it in some time - thanks for the suggestions, at first glance they look very appropraite

[Jekotia]

There IS a use case for container names: having human-readable container names in docker ps output, and any tools that let you view running containers on a docker host (e.g. Portainer, Komodo). Without defining this field, the container is given a random identifier for the name.

[vanodevium]

@Jekotia

Fortunately, the name is never random and is built according to very clear rules.

I only removed them because these names were used for DNS resolution

[Jekotia]

@vanodevium I'll have to test later then. My experience had previously been that the name is akin to a commit hash. Seemingly-random generated gibberish that corresponds to something internally. Now I'm left wondering if that got changed, or if I misunderstood what was happening before 🤔

[vanodevium]

@Jekotia

Most likely, you are a little confused about what names are assigned by docker.

When a container is started within a compose, the name is built according to approximately this rule:

(project or cwd) + service name + number of scale (default 1)