The list of IaaS alerts can include:
1. CPU utilization
2. Disk utilization
3. Memory utilization
4. Server response time
5. Time aggregation (a statistic that is run over raw metric values – supported time
aggregations are Min, Max, Avg, Total, Count Average)
Of course, your business SLA will cover what you want as human alerts and what you
want as automated fix action alerts.
Figure 4: Azure sample alerts
With IaaS, you have control over the alerting. The biggest decision you will have to make
is whether to extend your existing on-premises alerting tools into the cloud; to integrate
a third-party tool; or to use cloud-native tools offered by the cloud providers. There are
advantages to each.
Cloud Provider Tools
Each of the major public cloud providers – Amazon Web Services (AWS), Microsoft
Azure and Google Cloud Platform (GCP) – has its own specific cloud alerting features.
The names are different (Amazon Cloudwatch, Azure Monitor and Google Stackdriver)
but the functions are basically the same. The good thing is, each one provides a full slate
of alerts you can configure for your own purposes, for specific workloads or actions. The
downside is, they do not work with each other. You cannot mix and match alerts for each
cloud, so keeping track of which alerts align with which cloud provider can be
complicated.
40 | THE DOPPLER |
FALL 2019