Feedback for “Improving Performance”

[miparnisari]

Feedback for https://authzed.com/docs/spicedb/ops/performance

1. I haven't verified this for every single SpiceDB flag but it seems that the default value for each flag aim towards correctness & solving the new enemy problem rather than towards speed. We should verify if my statement is true, then if so, document that in this page AND in the SpiceDB CLI help output so that users conducting load tests are aware.
2. The current page talks about: configuring dispatching, enabling materialize, and enabling the schema cache. There are more flags that can be used to improve performance or when using large clusters in production. For example:

• revision quantitization window
• watchable schema cache
• connection pool size and max lifetime jitter
• tx overlap strategy
• dispatchChunkSize (maybe?)