filibuster-testing / filibuster-corpus Goto Github PK

Hi,

Great job with Filibuster and this corpus! Based on my fair share of experience with microservices, this corpus stands out particularly for the sheer number of development hours spent to put out something valuable as part of open-source.

I am working on an automated cloud resilience tool, Phoenix. I want to try out some experiments (to emulate real-world micro service applications) and test the applicability of my resilience tool. To give a little bit of background, we are envisioning Phoenix as a resilience manager that sits on top of cluster orchestrators such as K8s and turns off non-critical microservices "safely" for large scale disasters. The idea is to still be able to server large number of user requests with reduced functionality while cutting down on resources --- something like call graph pruning. My idea was to use filibuster's corpus in the following way:

1. Deploy these demo applications on K8s
2. Generate load on these applications
3. Delete nodes in K8s arbitrarily
4. Measure how quickly does Phoenix prunes the microservices and how much of the load is still being served with this reduced functionality "correctly"

Does this seem like a practical application of filibuster's corpus or do you think there are some gaps in my understanding about filibuster wherein I have to do a lot of engineering to realize the above scenario?