Feature: Execution client and consensus client failover

[yorickdowne]

Suggested feature: Comma-separated list of "execution client" (eth1) and "consensus client" (eth2) connections. Earlier in the list has higher priority. On node failure, the next one is tried, and so on; if a higher-priority node comes back online, it gets used instead of the lower-priority one.

This could also be handled with proxies instead on the infrastructure level.

Rationale: Node operators need a way to do maintenance, such as upgrading versions, pruning Geth DB, etc, without their SSV node going down because of it.

[antondlr]

adding to this (important, essential even for production environments, imo) feature request; I'd prefer for the client to retry-loop, instead of fail, when its connection to the BN (or EE) fails.

[chichi13]

Do you have any news for this feature?

I think it's important to have it!

[antondlr]

I would like to re-surface this feature request now that mainnet setups are being rolled out. IMO this is essential for a robust, high-performance stack!