Add continuous performance benchmarking to Cosmos

[jbandoro]

After performance integration tests were added in #827 a reasonable follow-up would be to integrate a tool like github-action-benchmark to have continuous benchmarking of Cosmos performance so that possible performance improvements or regressions by comparing benchmark results can be detected.

The work would involve using pytest-benchmark to output the performance benchmark results and then storing the results with the GH action like in the example here. We could setup alerts for PRs.

A follow-up to this could also involve benchmarking the DAG parsing times for various load methods, e.g. LoadMode.DBT_LS and LoadMode.DBT_MANIFEST to track performance for parsing.

[dosubot]

🤖

That sounds like a great idea! Continuous performance benchmarking can definitely help in identifying performance improvements or regressions. Using github-action-benchmark and pytest-benchmark seems like a good approach. Your proposal to also benchmark the DAG parsing times for various load methods is a valuable addition. We encourage you to go ahead and create a pull request for this feature. Your contribution will be much appreciated!

About Dosu

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.