Sometimes GCE doesn't have enough available resources to start a new VM or GKE cluster. It could be interesting to track this over time and report which regions run out when.
I'd probably start it as a GitHub Actions bot that runs hourly(?) and tries to create a micro VM in each region, and reports if it fails due to stockout. Add some kind of dashboard based on that that reports %age of time the region is full, and/or what times of day they become unavailable.
I'm not sure if we'd need to report zone-by-zone or just region-by-region.
Make sure all VMs get deleted quickly after they get created or they'll cost money.
This is really something GCP should report itself, but 🤷.
Sometimes GCE doesn't have enough available resources to start a new VM or GKE cluster. It could be interesting to track this over time and report which regions run out when.
I'd probably start it as a GitHub Actions bot that runs hourly(?) and tries to create a micro VM in each region, and reports if it fails due to stockout. Add some kind of dashboard based on that that reports %age of time the region is full, and/or what times of day they become unavailable.
I'm not sure if we'd need to report zone-by-zone or just region-by-region.
Make sure all VMs get deleted quickly after they get created or they'll cost money.
This is really something GCP should report itself, but 🤷.