Welcome to the fourth article in our "Mastering Cloud Cost Optimization" series. I’m joining you from ParkMyCloud, a Turbonomic company that focuses on cloud cost optimization – in particular, scheduled workload suspension, which is what we’ll focus on in this article.
You can find the previous articles in this series here:
- Mastering Cloud Cost Optimization: The Principles
- Mastering Cloud Cost Optimization: Frameworks for Success
- Mastering Cloud Cost Optimization: Cloud Cost Models & Discounts Overview
Why Scheduled Workload Suspension Is Important
Scheduled suspension of workloads means scheduled “off” times for your on-demand public cloud resources, including instances/VMs, databases, container nodes and clusters, and more. A common example is for non-production workloads, such as development, testing, staging, and QA; which can typically be suspended overnight and on weekends. By suspending a resource, say, 7 PM to 7 AM and on weekends, you reduce your monthly cost by 64%.
The reason scheduled suspension stands out as a cost-saving method is because it’s wasted spend. On demand resources in AWS, Azure, and Google Cloud are billed by the minute or second. As Mor discussed in our first article, the reality is that you pay for what you provision. That means that any time VMs and other resources are allocated and left running when not being used, you’re paying for time you don’t need. We estimate that $11 billion will be wasted on idle resources this year alone.
Given the amount of waste, and the relative simplicity on the cloud cost optimization maturity curve, scheduled suspension is one of the first tactics cloud customers should take action to address to optimize costs.
How to Identify Wasted Spend from Idle Resources
Identifying potential resources to be scheduled is not as easy as deciding that everything will be turned off at 7:00. With disparate teams, time zones, and applications, you will want to utilize input from end users, or more likely, the usage data from resource utilization history to determine when resources can be turned off.
You can attempt to do this manually, but at scale with 100’s or 1000’s of workloads, this is no longer feasible. Once accounting for all IaaS and PaaS resources across accounts, regions, and resource types, there is too much variability.
The public cloud providers all offer some rudimentary native capabilities. They are better than nothing, but are typically limited by resource type (e.g. on-demand compute only), do not provide usage-based recommendations for optimization, are not multi-cloud, require scripting, don’t have built-in RBAC and don’t integrate with your daily operational tools.
These shortcomings are why third-party platforms have developed to fill the gaps left open by the cloud providers. When evaluating these, there are several considerations to keep in mind.
Take Action – with This Checklist in Mind
Ultimately, scheduled suspension can be straightforward to address. Keep the following in mind:
- Provide immediate ROI – be able to validate the actions you are taking and demonstrate value through reporting.
- Data-driven – make decisions based on usage data that shows when resources are actually used and when they can be suspended. Otherwise, changes will be discarded and costs rise again.
- Automated – automate actions and integrate into existing processes like patching and CI/CD. Require as few disruptions to user behavior as possible.
- Simple implementation – time is limited and your team has many priorities, so keep your solution straightforward and easy to implement.
- Self-service access – users are distributed and need to maintain access to override schedules as necessary.
- Multi-cloud – even if you are using a single cloud provider today assume you will go multi-cloud.
The ParkMyCloud platform aims to address these needs, providing automated cost optimization for your public cloud resources through scheduled suspension and more. Try it free today.
Stay tuned for the next article in this blog series, “Mastering Cloud Cost Optimization: Rightsizing. The Right Way.”