Currently conductor needs to be started after a rollup node implementing execution API. This is bad for development flow as it doesn't allow hot restart etc of services. We should update this to do a couple of things. The first is a simpler implementation likely unblocks most workflows, the second allows for dev workflows with a hot reload type setup (ie I update my rollup binary and blockchain restarts seamlessly).

1. On initial startup, poll for execution layer to be ready. Log that it's not but do no restart.
2. If execution API dies, we can restart services internally. The Geth implementation now responds with a PermissionDenied if GetGenesisInfo and GetCommitmentState are not called before attempting block execution. This would also indicate a need to clear cache and restart internally. (This should be documented as required behaivior)