Currently, we sometimes get neonvm request failure metric labels like
[error: virtualmachines.vm.neon.tech "compute-rapid-wave-a45g5t85" not found]
Thankfully in practice the rates are very low, so it's not a big issue, but in theory we could accidentally blow up our metrics storage if some issue was causing our requests to fail for every object in a way that included the object name.
Currently, we sometimes get neonvm request failure metric labels like
Thankfully in practice the rates are very low, so it's not a big issue, but in theory we could accidentally blow up our metrics storage if some issue was causing our requests to fail for every object in a way that included the object name.
So this PR makes sure we filter out the name.