Tools Research - Pranav-SA/thesis-support-examples GitHub Wiki

Approach

Why looked into these?

Chaos engineering is not yet a segment of the market that is well-established and developed. Nevertheless, there are several tools we can pick from. We cannot use them all, it would take too much time, and might not solve the purpose. So, we’ll have to pick. Contributing Factors:

  • Open-source (Low running cost)
  • Work inside or outside Kubernetes
  • AWS or other provider's plugins
  • Manual executions or automation by scripting
  • SAP's first-hand experience or development let’s see which tools we have at our disposal

Chaos Monkey or Simian Army: There are Chaos Monkey, Simian Army, and other Netflix tools aimed at chaos engineering. What Netflix did with Chaos Monkey and the other tools is excellent. They were pioneers, at least among those that made their tools public. However, Chaos Monkey does not work well in Kubernetes. On top of that, it requires Spinnaker and MySQL.

Gremlin: It might be one of the best tools we have on the market. But it’s not open-source and such a service only makes sense when the product has matured.

PowerfulSeal: Further on, we have PowerfulSeal, which is immature and poorly documented. Besides that, it works only in Kubernetes and does not not have extensive drivers.

kube-monkey: We also have kube-monkey, which is inspired by Chaos Monkey but is designed for Kubernetes. Just like PowerfulSeal, it is immature and poorly documented.

LitmusChaos: Then, we have LitmusChaos, which suffers from similar problems. It is better documented than other tools mentioned, it can be a choice in the future. It is green, and it is Kubernetes only. Gloo Shot: It is developed by solo.io. Gloo Shot is relatively new, and it works only on the service mesh level only. So, it’s also not a good choice.

Chaos Toolkit: Finally, we have Chaos Toolkit. It is very well documented with modules that significantly extend its capabilities. We can use it with or without Kubernetes. We can run experiments against GCP, AWS, Azure, service mesh (Istio in particular), etc. It has a very active community.

image