How does CloudMonitor reduce your Azure bill and avoid Bill-Shock?
We are getting a lot of questions on how CloudMonitor.ai works and whether it is better for PaaS or IaaS – here is a common answer. CloudMonitor is constantly monitoring your Azure resources and Azure spend.
It works on a few different levels to help you reduce your Azure bill:
1) Cost Optimisation recommendations – in addition to the 17 Azure Advisor recommendations we add 50 more of our own custom rules. Simple things like looking at geo-replication settings on Storage Accounts, recommending an HDD disk instead of an SSD, comparing App Service metrics versus size, checking SQL or Cosmos over-sizing… etc etc. (there’s a number of rules for each Resource Type mainly around the concept of metrics versus size).
2) Bring accountability to the creators of the Resources. The Teams/Slack bot will tell you the next day if you have done something like leaving a SQL DW running with no load or left a Databricks cluster on Development without AutoShutdown enabled (we call these Cost Mishaps- currently the Bill Shock happens at the end of the month where the SysAdmin who has no idea what the lower-level resources are for, gets an expensive bill. It’s a huge effort to find out who created a resource. We also Auto-Tag the actual creator and Created Time of the Resource in Azure as it happens – something Azure is lacking. With Accountability, you can now start to identify WHO is responsible for the most expensive resources.
3) Cost Anomalies – we use ML to let you know the next day if something is out of the ordinary from the previous day trend. This also avoids bills shock by shortening the feedback loop.
4) We have a new view called CostGroup which is not technical like a Resource Group or Subscription – it is a logical grouping of resources by business function (some people use Tagging to achieve something similar, but we provide this natively).
5) There are pre-canned Power BI reports for CostGroup owners. E.g. a PM will get the cost summary for their projects every month and the Execs will get a high-level summary of costs. Developers get a list of the resources they created and how much they are spending and the recommendations for those resources. You can also do your own analysis with Power BI which some find more flexible than the Cost Management tools.
6) Remediation action – CloudMonitor can automatically fix the scale issues if you want it to, or you can do it yourself. Some people don’t like the idea of automated cost governance on Prod (fair enough).
7) Scheduling (coming soon) – a lot of people use automation scripts to pause resources that require custom code and effort. CloudMonitor lets you make schedules to match your business hours and apps and assign it at the CostGroup level by Environment – so that any resource (PaaS or IaaS) within that CostGroup will be started or stopped according to the schedule. You don’t always need the DEV environment running and a lot of custom scripts don’t cater for Public Holidays etc. Did you know that normal Business Hours are only 35% of a 7 day week? Are your services on for the other 65% of the week?
To answer your question, we focus mainly on PaaS, however, there are many gains to be had with IaaS systems with the scheduling functionality.
Your mileage may vary, but initial tests have been very encouraging, especially in bigger organisations where governance is hard to implement.
Let us know if you have any more questions on how it works. We’re also looking for more BETA testers – in exchange for your feedback we’ll give you a 1-year license and aim to reduce your Azure bill by 25% for free! Contact us to discuss…