Profiling performance and tracing execution and memory bottlenecks #2600
Description
Profiling and tracing bottlenecks related to app -> extension overhead, and extension -> indexeddb overhead, and (2) minimizing RTT / latencies using vitest profiling. A shared excel sheet tracking this information should be linked here for reference.
Reducing view service latency depends on optimizing storage accesses, which consistently suffer from O(N) scanning complexity. This likely requires storage migrations to IndexedDB, and we should profile to understand if this is a bottleneck vector. We haven't explicitly benchmarked linear scans, but i'd posit that it's reasonable to infer they're a contributing factor to the bottleneck. Also, any place where we might be serializing network latency (I don't think we're doing that), we should kick off the rpc calls concurrently and await the promises in parallel.