2 ms·

Or just deploy containers with an orchestrator restarting them when failing? It is not like an Erlang service would be able to make progress with an invalid co

Or just deploy containers with an orchestrator restarting them when failing? It is not like an Erlang service would be able to make progress with an invalid config either.

lenkite 11mo ago
What happens when they "keep" failing ? You never get to know what is causing your nightmare.
jacquesm 11mo ago
That's fair, but even there the roll-back would be a lot smoother, besides the supervisor trees are a lot more fine grained than restarting entire containers when they fail.