Timely Health Alert reporting
Difficult to find the best place for this....
Today at 09:45 GMT, I discovered that I could not start a de-allocated VM in UKSouth, and raised a Microsoft support ticket at 09:47BST.
No issues were shown on the Portal Service Health and no alerts were received, until 12:32 -apparently the fault started at 06:50 UTC.
And then the Portal Service Health showed the issue at UKWest and not UKSouth.
As a consequence this has got me very concerned regarding running a service and reporting to my customers, and seems at odds with Service Management Reporting at SLAs.
I am less concerned that I couldn't start my VM, that what seems like poor and late service reporting.
Please reassure me that this is a one off....
It's 15;33 and still showing as an issue
Jean-Philippe Page commented
I'm not sure if this is the same thing but yeah, I am very worried about the time it takes before alerts are sent out. I'm in the process of automating our alerts and also doing some testing on them. I set an alert to tell me of any administrative actions done to a MySQL server. To be sure the alert was properly activated I waited till the next day before I started testing. I then changed a setting on the server and checked the alert section. It told me the alert was fired... but it took 50 minutes before I received an email for it.
Is it also this long for metric alerts? I can't wait 50 minutes before finding out that something is having issues, this makes alerts completely useless and unacceptable.
Mike Binnix commented
This must be a high priority. We had an even more extreme example where customer service reported a systemic network failure in both US East2 and UK South. We could not start or provision any VMs for a 29 hour period during the service impact. BUT THE EVENT WAS NEVER REPORTED ON THE PORTAL SERIVCE HEALTH for US East2. NEVER MENTIONED. I find this an unacceptable omission in the process for managing system outages. The lack of communication is worse than the service disruption itself.