How do folks decide when to use a native | Dagger | Page 1

weak finch Jul 5, 2024, 7:28 PM

#

There are pros and cons to each approach, but I think a lot of it comes down to caching.

For now, one rule of thumb is that Dagger will automatically cache a bunch of stuff that happens inside of DAG that is generated by the Dagger Functions you're using like WithExec, but not the parts that are written in standard language code.

tidal fiber Jul 5, 2024, 11:14 PM

#

Ahh gotcha that makes sense. If I use standard language code, is the compilation of that still cached at least? This makes me think that in general I might be better to take my standard language code and have dagger compile it and put it into a container and then WithExec it that way.

#

I like the ease with which I can execute a CLI command like aws without having to setup all the auth/client stuff myself if I’m doing something simple like retrieving an ecr token.

weak finch Jul 6, 2024, 5:06 AM

#

This is where you would want to look into using cache volumes, to cache things like node, go packages, or python packages https://docs.dagger.io/manuals/developer/cache-volumes/

Cache Volumes | Dagger

One of Dagger's most powerful features is its ability to cache data across pipeline runs.

#How do folks decide when to use a native