This is the first in our series of habits described in our article5 Habits of successful Cloud Cost on a diet.
Habit #1 Set the guardrails to kick off the financial (RI & SP) management on Autopilot
The fastest way to reduce your costs operationally in the cloud infrastructure are with our Habit #1, where you set the guardrails and put the financial management of your cloud on Autopilot, often reducing your bill by 11% or more.
The best part about these savings :
- High & immediate savings on your monthly bill
- No operational or engineering impacts
- Low risk
As an added bonus, when your coverage true-up happens on a weekly cadence to the approved goals, it eliminates the manual operational overhead needed for large quarterly/yearly purchases.
First step is to answer the following questions:
- Do you know your current average reserved instances coverage for your EC2, fargate and lambda workloads?
- Have you established a goal for reserved instances coverage, and is it higher than 85%?
- How much 1 year and 3 year commitments should you make?
- Is there any waste from existing Saving Instruments? Is the utilization of reserved instances over 95%?
- How much capital if any you want to invest upfront for better savings rates?
Define your goals as if it were your desired weight. How much do you want to save with savings instruments? Whichcombination of savings instruments fits your requirements?
Next step is to invest in some tooling to automate the manual weekly process and build the capability to perform weekly true ups. Cloudwiry delivers this with our expert monitored autopilot that our customers have been relying on now for 3 years.
To get maximum savings, you would want to purchase 3 year savings instruments at full coverage, maintained weekly. Running at 100% on-going coverage with a rolling 3 year commitment is often perceived to be high risk due to future uncertainties.
Good news is that this risk can be effectively mitigated uniquely in AWS with Convertible Reserved Instance as discussed in my blog onRebalancing your convertible RIs to reduce your monthly commitment and weekly true up of existing Convertible Reserved Instances. Weekly true up of convertible RIs does not increase your term commitment, thereby further improving your hedge.
For instance, we had a customer who increased coverage goals from 55% to 85% coverage and were considering purchasing 1-year Savings Plans for additional coverage needs.They had existing 3 year convertible reserved instances with ~360 days-left instead. Performing upsizing operation of these convertible reserved instances by increasing their contract size, for the same commitment (~1 year) they gained an average increase of 23% on savings. Indeed, all the fleet of CRIs were upgraded from 1-year non-upfront to 3-years non-upfront.
In short, for each $1M/year of EC2 costs, finding the best approach enabled getting an extra saving of $231k/year, with no extra-commitment.
This is the first one of a series of posts, where we describe how to start optimizing the costs with the saving instruments provided by AWS. In the next ones, we’ll cover how you establish other guardrails and set up other automations to achieve your cost-saving goals. At a high-level, you should be asking your organization:
- To set up a plan for long-term financial success to reap the benefits the cloud promises to deliver.
- To tag specific areas/critical workloads/infrastructure/etc…to identify near-term cost savings and match this with your long-term plan to determine the owners of the infrastructure.
- To set up a path to automate mundane tasks and mitigate financial risks.
- To handle implementation, daily checks, and establish the right cadence needed.
- The hard stuff: you wouldn’t want to miss this. Get up to the last 15% of coverage.