Insights

Either using blood tests, medical imaging, questionnaires, or simply patients describing their symptoms, doctors initially conduct a diagnostic to identify potential health issues. Similarly, in the business world cloud specialists and consultants conduct diagnostics to find potential financial burden and saving opportunities for their clients.

A Cloud diagnostic is defined as is the initial cost optimisation assessment to present an overall and a high-level cost saving opportunity. It is conduced within a brief period of time, and it is initiated by sending a questionnaire to our client to better understand their company, platforms, demand signals and cloud cost growth. 

Saving opportunities are found within different categories including: compute, storage, backup, recovery, and others such as network and security. From the Cloud diagnostic engagements that I have been involved with, I have noticed common cost savings opportunities/areas shared across different clients, including:

1. Low CPU utilisations: Companies tends to over provision their resources, leading to instances with a very low CPU utilisation, causing the clients to pay for unused resources. To overcome this, we advise our clients to right size their resources by changing their instances to a smaller and cheaper instance or to turning off instances that are not used.
   
   
2. Instances generation: AWS is now offering new generation instances with the same number of cores and memory but lower in price compared to old generation instances (saving could potentially be up to 10% depending on the instance type and Availability Zone).
   
   
3. Low memory usage: Similar to CPU utilisation, memory is sometimes underutilised. We recommend switching to a smaller instance types.
   
   
4. Demand signal vs cloud cost: In an ideal world we would expect cloud costs and demand signals to be proportional, with a similar growth pattern. If this is not the case, investigations are required to understand where the additional growth comes from.
   
   
5. Payment plans on demand vs compute savings: Amazon offers less costly payment plans to reduce EC2 cost, with up to 66% off with Compute Savings Plans compared to on demand payment plans.
   
   
6. EBS volumes: Customers would need to have larger gp2 volumes to gain higher performance for some applications, such as MySQL, which require high-performance space. However, in some circumstances, those applications do not require significant storage, resulting in over-provisioning of storage or performance, which is not cost-effective. We therefore advice our clients to migrate their gp2 volumes to gp3 volumes, which offer regardless of volume size, 3,000 IOPS baseline performance and 125 MiB/s for free. As a result, clients can provision smaller volumes at a lower cost while still ensuring exceptional performance.

As you can notice, these opportunities go beyond just analysing cloud cost optimisation. They require optimisations across different areas of the cloud infrastructure such as capacity, performance engineering, cloud providers and your own teams. Some organisations decide to sacrifice one area for another, for example, increasing cloud costs for better Capacity and performance.

In Capacitas experience, we have achieved implementations without sacrificing one area for another. We have proven time and time again that our clients can improve performance, reduce costs, and optimise for capacity simultaneously while improving the capabilities of your teams.