Can I abuse the cache this way?

I’ve built some code into my project that runs benchmarks and saves the data (in a .csv file) and generates a couple of charts, to show historical trends.

It’s easy enough to run this with some tweaks to circle.yml.

However, I really want the changes to the .csv file to persist build to build.

Is this something I can do by adding the directory containing the .csv file as a cache directory?

Reviewing the docs, I can see that ache is saved before the test phase, which is wrong for me. I think I need an entirely new concept to make this work.

The caching feature doesn’t sound like something you should use. What kind of data are you working with exactly? Our Test Metadata might be useful to you.

Aside from that, I would consider saving these CSV files in the one of the following locations:

  • in the git repo itself
  • in another git repo
  • AWS S3 bucket
  • metrics SaaS such as Datadog