Open faddat opened 2 years ago
I love the idea, and am frustrated by state-sync's reliability as well. We're planning on doing some work to alleviate the state-sync pains, and it'd be great to have this monitoring in place before that.
That said, I don't think it should be in this repo. I'll make a network-status
repo for this and other monitors, which we can run on a cron schedule. That work?
I agree with @boz. If this is just going to be an effort to proactively test the health of the software & the network, maybe we should create another repo? It can live independently of this. We should not be redesigning or reworking state-sync each update. So the test would stay close to the same without needing much in the way of adjustment each update.
That makes sense and I'm happy to add the former PR over there.
Checked with @troian , we still want this except that the cronjob should be somewhere else; not the node repo.
as per Artur during today's sig-support meeting:
mainnet
;GH action within node repo and can be triggered either manually or on release candidates
. Or use a matrix state.
Is your feature request related to a problem? Please describe. State sync is cool, but unreliable.
Describe the solution you'd like Add a github action that tests state sync with each software change. Make it a first class citizen.
Describe alternatives you've considered I have tried many ways to sync nodes, and they are all much too slow and much too stateful.
Additional context akash-network/node#1366 is a first stab at this. It is not expected to pass all the time, but instead to give us an idea of how it's going.