Consider supporting mocking Cassandra/astra and fault/latency injecting #119
Description
In our testing we have a process where we test our application's binary as it goes to production. For HTTP we use contracts to stub data and we use fault and latency injection to mimic production behaviour. Currently we use either an actual single Cassandra node, or a full cluster when the application needs Cassandra. This has the benefit of testing the actual Cassandra part at the cost of the runtime. The other drawback is that there is an actual state, so a DELETE may break the next SELECT.
Looking at the mockcluster code there seems room to make queries/statements read-only.
Is it within the scope of this project to be either or/and:
- A proxy to an actual CQL/Cassandra cluster with fault and latency injection.
- Supporting fault and latency injection on top of an in-memory mock database
I would be willing to contribute some time on this if I get some pointers on how best to implement this in line with the project.