refactor: improve failover provider

[jtourkos]

Hey! This implements events-processor#37

I've made the same changes to app, multiplayer and api projects. Let's review the changes here first and I'll apply the feedback to all the rest, keeping those PRs focused just on integrating this updated version.

Two main things we fixed:

1. We now have proper logging so we can actually see what's happening with our providers - which one is being used, why it failed, if we switched to another one, etc.
2. Fixed a bug where we weren't handling provider errors properly. We only checked for HTTP errors before, but a provider being "down" could also mean getting a 200 response with an RPC error in it. We handle both cases now.

Once this looks good, we should finally see clear logs telling us when we switch from Infura to Alchemy and why. This should help us understand what's actually happening in prod.

[railway-app]

🚅 Previously deployed to Railway in the Drips App project. Environment has been deleted.