invalid memory address or nil pointer dereference azure.mapFromVMScaleSetVM about prometheus

[darshanime]

from the stacktrace, the issue seems to be here. let me raise a PR

 func mapFromVMScaleSetVM(vm armcompute.VirtualMachineScaleSetVM, scaleSetName string) virtualMachine {
-	osType := string(*vm.Properties.StorageProfile.OSDisk.OSType)
+	osType := getOSTypeFromVirtualMachineScaleSetVM(vm)
 	tags := map[string]*string{}
 	networkInterfaces := []string{}
 	var computerName string
@@ -702,3 +702,23 @@ func (d *Discovery) getFromCache(nicID string) (*armnetwork.Interface, bool) {
 	net, found := d.cache.Get(nicID)
 	return net, found
 }
+
+func getOSTypeFromVirtualMachineScaleSetVM(vm armcompute.VirtualMachineScaleSetVM) string {
+	if vm.Properties == nil {
+		return ""
+	}
+
+	if vm.Properties.StorageProfile == nil {
+		return ""
+	}
+
+	if vm.Properties.StorageProfile.OSDisk == nil {
+		return ""
+	}
+
+	if vm.Properties.StorageProfile.OSDisk.OSType == nil {
+		return ""
+	}
+
+	return string(*vm.Properties.StorageProfile.OSDisk.OSType)
+}

[McHammar]

If the azure service discovery finds a VM scale set then we get this error. If the service discovery is for a subscription without one it still works

[bboreham]

Closed by #13578

[dpericaxon]

Hello I just updated to v2.50.1 from v2.49.1 and started seeing an error that looked similar:

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x123e50a]

goroutine 1605 [running]:
github.com/prometheus/prometheus/discovery/azure.(*Discovery).refresh.func1({{0xc180dc21e0, 0x92}, {0xc180db2de0, 0xe}, {0xc180db3362, 0xe}, {0xc179d65620, 0x21}, {0xc180db2df0, 0xd}, ...})
	github.com/prometheus/prometheus/discovery/azure/azure.go:438 +0xb4a
created by github.com/prometheus/prometheus/discovery/azure.(*Discovery).refresh in goroutine 1174
	github.com/prometheus/prometheus/discovery/azure/azure.go:375 +0xa55

Is it possibly related? Has anyone else thats updated seen this?

[bboreham]

I think it's more like #13643, which was a bad merge, fixed by #13702.

There is a release candidate for 2.51.0 out now which has this fix.

[dpericaxon]

I think it's more like #13643, which was a bad merge, fixed by #13702.

There is a release candidate for 2.51.0 out now which has this fix.

Great! Will there be an official release for 2.51.0 today/tomorrow? Thank you for the quick response by the way!

[bboreham]

We need to wait longer than 1 day for feedback to come in on the release candidate.
Please try it out! If nobody has any feedback this then will become 2.51.0, you just get it faster.