Sort:  

We are currently not doing that publicly but if you run a project that needs this, we may consider it. We already have health checks and monitoring available, but that won't protect us from downtime if steemd has hickups.

no one is protected from steemd hickups;)
but at least maintenance cycles, so ops can switch from one node to another, during that time window.
Like a real service :))

In that case I recommend backup nodes if your library allows them. It can happen that nodes just lock up and just don't respond at all without warning. No one can predict that sadly.

Define stable ...

api.steemit.com is stable enough - otherwise use backup-nodes

define stable enough ;)

Also remember that we provide an rpc node without jussi or appbase or other beta features that made problems in the past. We may change that once we deem them stable enough but for the moment we have users who depend on us not changing node behavior on a whim.

@reggaemuffin

Your RPC-Node is top-notch. Don't get me wrong there. I just believe api.steemit.com is stable enough for most use-cases and in combination with backup-nodes, there shouldn't be any problem there.

Yeah definitely! For small use case they are enough. But once you leave that territory you can be pretty left alone should issues arise (like the current one where api.steemit.com lacks behind on transfer history by days)

Oh, that timing is precious :)

they switched to version 0.19.4 without any notice.
This is a no go in any service management.
You have deployment plans, rfcs, etc.
I should write a tutorial for steemit it seems, how to operate a reliable service.