Orestis Floros

> To achieve our objective of categorizing error messages into different types for effective monitoring - distinguishing critical issues (like high resource usage, restarts, memory problems, etc.) from expected errors...

Or perhaps even: ```diff diff --git a/resources/providers/awslib/s3/provider.go b/resources/providers/awslib/s3/provider.go index add35694..021078bd 100644 --- a/resources/providers/awslib/s3/provider.go +++ b/resources/providers/awslib/s3/provider.go @@ -79,26 +79,32 @@ func (p Provider) DescribeBuckets(ctx context.Context) ([]awslib.AwsResource, er bucketsRegionsMapping := p.getBucketsRegionMapping(ctx, clientBuckets.Buckets)...

Note: poetry has a "CI recommendations" section on its website: https://python-poetry.org/docs/#ci-recommendations Additionally, it might be quite challenging to add a hermit package for poetry as hermit packages mostly just describe...

I wonder how easy it is to have our own all-in-one github action that roughly does the following: 1. Checkout repo 2. Set up hermit 3. Set up go +...

## Suggestion for safer channel usage in fetchers This is a toy implementation of my suggestion in the retro. As a demo, I replaced the channel in the IAM fetcher...

@olegsu writing on a closed channel is a panic and we can't do anything about it, your program exits with a panic. This is why writers should always be responsible...

@jeniawhite @olegsu After yesterday's discussion, here's a toy project that illustrates how we can work with multiple workers that can be independently restarted with separate contexts + safe context-aware channel...

Hey @amirbenun I am not sure the implementation in `pipeline.go` is similar, since it doesn't handle the context done case, we will have to fix that usage as well. The...

I am re-tagging as enhancement since this is intentionally done this way and it is something we can improve on a future version by weighting the additional onboarding, configuration and...

> One way of doing that is to modify the fetcher interface to return a receive-only channel. But this means creating one channel on each fetching cycle?