Restart does not have limit

[shmel1k]

Hello.

By now, if I'm trying to restart huge clusters (for example, 1000+ storage nodes), I can get funny consequences: overload, statestorage ddos and something like that.

The idea is to limit parallel restarts and add pauses between restarts

[kobzonega]

For example some useful tools in my company have args like --delay-between-restarts and --ssh-pool-size which implement this logic