Reading a FileChannel that does not have a writer blocks
indefinitely. This occurs in cannon whenever the pre-image oracle
server shuts down prematurely. We add a new io.ReadWriter that can
detect when a FileChannel would otherwise block indefinitely.

fix(indexer): e2e test lifecycle

op-challenger: Use gzip encoded proofs and snapshots

Add ability to restrict the networks that op-node discovers peers on

op-challenger: Include the op-program host binary in docker image.

op-challenger: Detect already resolved games at startup

Update rollup-node-p2p.md

Avoids the need to manually specify the op-program binary to run as the cannon pre-image server.

Typo fix

Avoids regenerating the cannon traces when performing an initial move at startup if the game is already resolved.

ci: Convert go-e2e-tests CI to make targets

This change makes it much easier to make build changes to the op-e2e
without having to reflect the changes from the Makefile into the CI
definition.  It also makes easier for developers to mirror the execution
of what's going on in CI locally.

feat(indexer): wire up ETL metrics

feat(indexer): introduce metrics server for the indexer process

chore(indexer): Use middleware for healthz

op-challenger: Separate multigame and single game code into different…

fix(indexer): application-level context. better go-routine management

op-challenger: Add CLI option for max concurrency

Add note about arbitrary message sending from OptimismPortal