Resonate Recipes: Basic Failure Handling
In this edition of Resonate Recipes we will explore catchable and non-catchable failure and how Resonate mitigates these failure types.
To follow along, head over to the Resonate Recipes repository on GitHub, clone the repository, and navigate to the Basic Failure Handling recipe.
Catchable failures refer to the set of failures that can be detected and mitigated by a process itself e.g. in a try catch block. Examples include io failure such as a file not found exception or networking failure such as a request timeout exception.
Non-catchable failures refer to the set of failures that cannot be detected and mitigated by a process. My favourite mental model is to imagine the plug being pulled on the machine running a process.
Resonate detects and mitigates catchable failure via transparent try catch and by retrying executions. To see this in action, try running the following program. If you have cloned the recipes repo, you can run this example by running npm run catchable.
Running this function directly would, on average, fail 50% of the time. But executing via Resonate (almost) always succeeds. Why is this?
