Manu Garg

I think the state that @alexdowad is talking about is when the probe was run last -- this makes sense for the setups that run probes infrequently, say with the...

It will be a slight bigger change for surfacers as I'll have to first move surfacers/internals.

No worries. I am attempting it in: https://github.com/cloudprober/cloudprober/pull/1188 Just want to make sure most used interfaces are not broken.

#1188 is merged. Please give that a try. Here is the exported interface: https://github.com/cloudprober/cloudprober/blob/f6d8378c68feacb92434ab9a02e6e49f6f4bc9ee/pkg/protos/surfacers/surfacers.go

Hello @thomas-desforges, do these ingresses have address assigned while still getting provisioned? Ideally ingress status should change to include the address only once provisioning is done.

Hi @dbednarski, sorry Firefox has so few users (just 63 right now), that it's hard to build support for firefox specific issues :( .. Sorry :(

This timeout seems to be coming from delay in probe server (`my_probe` in this case) responding to cloudprober -- it seems probe server is taking longer to respond from time...

Fields that we should verify: - interval - timeout - latency_unit - interval and timeout correctness I guess we should just run a limited version of: https://github.com/cloudprober/cloudprober/blob/01180b7a2f86fdbb9dac15987ead400ed110f0a5/probes/options/options.go#L150

@PovilasV1 Unfortunately merging metrics like this is problematic as both EventMetrics have different types (gauge vs cumulative). What problem are you trying to solve? Most surfacers don't care about metrics...

I see. One thing we can do is to make sure main metrics and payload metrics have same timestamp: https://github.com/cloudprober/cloudprober/pull/1163 That will likely not solve your problem, but probably provide...