When things are running smoothly one may not pay too much attention to monitoring. It’s when everything is on fire or applications break that we begin to work backwards to find the root cause of an issue. It’s from this scenario that we begin to talk about being “proactive” resolving issues before they happen, or “reactive” where find ourselves in break-fix mode. This also leads to proactive and reactive monitoring and alerts. Having a good monitoring and alerting plan can save you hours, sometimes even days to troubleshoot a problem and it can also help you to gain a more accurate view of how your infrastructure and applications are actually performing. Azure offers plenty of robust tools with tons of new features being released every month for monitoring and alerting.

In the illustration we can see that several metrics are created from our Azure VMs. These metrics include such things such as CPU utilization, network and disk I/O. This is made possible by the Azure VM Agent that runs on each newly deployed Azure VM. However, we also have the ability to turn on VM Diagnostics which gives us an even deeper view into our VM performance. Enabling Diagnostics on a VM deploys the AzureRmVMDiagnosticsExtension which gathers a wealth of useful information for monitoring including Event logs and stores them in your VM storage account. But of course with more data means more storage, so just be mindful of the amount of diagnostic data you’re using in your storage accounts.

Alerting is also available in Azure. We can configure alerts based on cpu utilization and other metrics and email service owners based on preconfigured thresholds.