More robust error handling in case anvil crashes

[fubhy]

We are currently tracking whether anvil fails to start, but after that, we don't really monitor the underlying anvil process anymore to capture any crashes or exits.

[tmm]

Inspired by https://github.com/wagmi-dev/viem/actions/runs/4863340150/jobs/8670981663?pr=440#step:4:2955 😃

[0xOlias]

+1. I'm running into a similar error.

FWIW, in my case I'm nearly certain the root error is that the RPC provider used for the FORK_URL is rate limiting requests (when I switch from Quicknode to Alchemy, everything runs smoothly).

[tmm]

@0xOlias Maybe helpful: The error I posted above came from a new anvil release so we pinned it.

[matthewlilley]

Yeah, this would definitely be useful.