maybe pool eth ws clients or create new ones ad-hoc?

when looking at the debug logs, I'm noticing that the service_request id is incrementing serially. That tells me that the same client is being used for all requests, which makes sense bc that's how I designed it. But I wonder if it will cause problems, i.e. if the ID maxes out its type -- that would practically never happen with a u64, though. What other problems might occur from using the same client for every request?
Does it force the responses to be processed serially by the node? In that case we could be missing out on parallelization gains.

No files, tests, or entry points are named. Start by tracing how the shared Ethereum WebSocket client generates service_request IDs and how responses are processed, then compare that behavior with pooled or ad-hoc clients; done means documenting whether reuse limits concurrency or creates other problems and choosing an approach.

Written by the indexing model from the issue text.