fix: Compare local nonce w/ backend to detect invalidated cache

[ryanschneider]

📝 Summary

If the user submits a tx via some pathway other than our rpc-endpoint (e.g. the public mempool or another private pool) then our cache of the users max nonce was not being invalidated.

⛱ Motivation and Context

Steps to reproduce:

• Send a tx via rpc-endpoint.
• This causes the nonce to be stored in redis for 24h
• Send another tx via the public mempool
• Wait for this mempool tx to be mined
• Query rpc-endpoint via signed eth_getTransactionCount RPC
• Note that the old value is still returned.

The fix:

Before returning nonce from cache, we still do a backend request as normal, if it returns a tx count greater than the one derived from our cache we return it instead and invalidate our cache.

📚 References

---

✅ I have run these commands

• make lint
• make test
• go mod tidy