Diagnostics and Monitoring
-
Add CPU and Memory usage metrics per instance
We use New Relic to monitor our cloud services. Sometimes we see an instance going above 80% CPU, which would possibly be solved by rebooting the instance, but it's imposible to identify which of the instances is in trouble in the portal (Azure portal uses _N to name the instances, New Relic uses an ID).
82 votesGuillermo, you can view the instance level data for CPU, memory and other metrics by turning on ‘Verbose’ level in the ‘Configure’ tab on the portal (https://manage.windowsazure.com). Please note that this is currently supported for Production servers only; support for Staging is coming soon.
- Ashwin Kamath, PM, Azure Insights -
Enhance Audit Logs on new portal with Timestamp and Started status
Old management portal shows Audit Log in pretty good way: Each record has detailed timestamps with seconds. It's absolutely clear when operation is Started and when it is finished (Succeeded/Failed).
Please consider enhancing Audit Logs on new portal with this information. It is very useful during troubleshooting.
14 votesThis is available in the Azure Portal. To configure which columns are shown for the Activity Log, simply use the “columns” button at the top of the Activity Log blade.
Thanks,
John Kemnetz
Program Manager, Azure Monitor -
Raise the Alert Monitor Limit per Azure Subscription
Currently, there is a ten alert limit for cloud services per subscription in Azure. This is functionally useless. We do four deployments per app for geo-redundancy, and if we want to monitor three different things on all our boxes, we're already over the limit. And we have dozens of apps.
Also, if you do increase the alert count, considering enabling an option to change monitoring level to verbose using the API, so we can automate the creation of monitors for memory usage.
8 votesAlerts will soon go GA at which point the limit will be removed.
-
Provide event streaming API on the Diagnostics Monitor
I need more real-time monitoring than currently is possible. The DiagnosticsMonitor should stream the events it captures to an attached listner.
We can then create a custom listner which registers with the diagnotics monitor for all or some specific events we are interested in.
we are then, code wise, free to transport that info to the storage container of our choosing.
It also opens up the possibilities of online service monitoring sites to provide a custom listner which you as a customer can deploy with your app.
This also has the benefit that you don't have to provide keys to…7 votesEarlier this year we released the ability to stream logs from Windows Azure Websites. However this capability does not yet exist for other services (such as Cloud Services).
-
mdsd
Hi,
as of now the MDSD (mandatory diagnostics service daemon) is only provided as a binary. There is no source code nor documentation of the protocols spoken available.
The MDSD binary provided is not suitable to be used on linux distributions other than the officially supported distros.
Withouth MDSD it is not possible to use the autoscaling features of a VMSS.
Thanks!
5 votesMdsd for Linux is planned to be open-sourced.
Thanks,
John Kemnetz
Program Manager, Azure Monitor -
Incorporate old portals management services Alert rules into new portal
It appears that I can't access the alert rules I defined in the old portal within the new preview portal... are there plans to surface these rules in the new portal?
3 votesThis will be addressed soon — any alerts on Virtual Machines or Websites will show up in the new portal.
- Don't see your idea?