AWS Knowledge
How to Analyzing Your EC2 RI Coverage Report?
Piyush Kalra
Sep 10, 2024
Have you ever thought that you could save a lot on the AWS costs, but don’t know how to? Or, have you bought a reserved instance, but want to make sure it’s not going to waste? The simplest way to do this is by going to the Coverage Report, on the AWS Management Console.
Starting from the basics… the most common way to optimize your expenses in the cloud is by looking into how you spend on compute. Do you use any Amazon EC2 Instances? If so, are you using any of the savings instruments like EC2 Reserved Instances, Spot Instances or Compute Savings Plans. But how do you know that you are making the best decision when it comes to your compute? How can you see 2 months later that it’s working? This guide will walk you through two different things:
1/ How to identify which instances should be reserved.
2/ How to tell if the EC2 Reserved Instance(s) you purchased is going to waste.
Understanding EC2 Reserved Instances
Before going further to the RI Coverage Report let us first understand what EC2 Reserved Instances are and their significance. You can reserve Amazon Web Services EC2 instances at a lower price than the On-Demand (paying hour-by-hour) price for a certain period of time (subscribing to an instance for a longer period of time – but still paying per month). This is because by committing to AWS for 12 or 36 months to use this instance, it guarantees to AWS revenue, so in return, they will offer a discount, versus the hour-by-hour model. The nice thing about reserved instances, is you also can choose “No Upfront”, which means you still pay per month, not all one payment. Think of it like renting an RV. If you rent it, it may be more expensive than buying it if you continue to rent the RV for several years / months. RIs are the same. Therefore, for a certain commitment period, you can expect major cost savings by improving on a specific instance type.
The Purpose of the RI Coverage Report
There are several dashboard views that are generated in the AWS console to serve as tools to understand which instances are covered by RIs, and which instances are not being covered (but could be), and which ones are being utilized. These reports can support you in pinpointing where you can optimize spend and save some money—think of it as a cloud costs fitness tracker! With these reports, you can manage your purchases sensibly by avoiding unnecessary / the wrong RIs purchases and still making the most out of every dollar.
There are two parts of the Coverage Report:
1/ Reservations coverage graph.
2/ Reservations coverage breakdown.
How to Read the RI Coverage Graph
A good understanding of how to read the RI Coverage Report is necessary for an effective evaluation of this report. These metrics not only state the performance of your current resources but also allows you to take the necessary steps toward ensuring cost savings. Here’s how to read it:
Average Coverage (%)
If you don’t have any reserved instances, then this will be 0%. If you have any RIs, then you’ll (ideally) have more than 0% covered. In other words, the Average Coverage percentage applies to the proportion of all the running instance hours that have been paid using AWS Reserved Instances (RIs). For example, if your total running instance hours are 1,000 and hours that had RI covered is 800, Average Coverage would be 80%. Usually higher percentage is indicative of more efficient use of RIs, which as a rule means that you have managed to take advantage of the RIs cost savings. To get to 100% Average Coverage, you would have a reserved instance for ALL instances you currently have.
The goal for most companies should be around 80% average coverage, as long as you have a stable infrastructure, and consistent spend, and believe those instances you currently are using are NOT going away for the next 1 year. Some of our customers choose to reach 85%, 90%, or 95% if they believe.
Total On-Demand Hours
This metric indicates the total number of instance hours that were put on On-Demand billing rate. If you had 300 On-Demand hours in a month, it implies other RI purchases could be done to optimize the costs, to close the number from 300 to 0. Analyzing this figure, you relate it to the times of usage and redistribute RIs towards a low spending season when the high spike in demand approaches.
Total Reserved Hours
Total Reserved Hours means instance Total Reserved Hours that will have been paid for using RIs. For example, if 700 hours are reserved and additional 300 hours are on demand but since 700 hours will enjoy RI discount, then one can plan how RH will be wasted. This helps in assessing whether you will have to reconsider the amounts spent on RIs in order to enhance the savings based on actual usage.
Total Running Hours
Total Running Hours is a cumulative of both On-Demand hours as well as Reserved instance hours. In this case with 300 On-Demand hours and 700 Reserved hours, Total Running Hours will be 1000. By reviewing this performance indicator against your On-Demand and Reserved hours, you can have a more comprehensive view of how well your resources have been utilized and hence strategize better with regard to the cloud infrastructure.
Total On-Demand Costs
This particular metric reflects the total cost incurred for the usage of On-Demand instance accessor. If for example, said costs for On-Demand for the month came to $600, then one of such costs is worth monitoring because it helps you understand how much RI cost coverage to increase in order to achieve savings or how to better utilize your instance. Monitoring these kinds of expenses helps you adopt measures which will cut down on total cloud costs without negatively affecting the required operational performance levels.
Learning these metrics would equip your decisions with facts to achieve better management of cloud resources resulting in higher efficiency and cost reduction in various operations.
Here, we can see a list of instances that an account is using for EC2. The two instances that are NOT covered are the m6i.large and c6a.large. You can see how many “On-demand hours” there are, and make the decision if you want to reserve these instances.
Analyzing the RI Coverage Report
Having the objective of determination of key metrics in hand, an approach toward the RI Coverage Report would now be provided which would serve as an effective way of reporting RIs.
Step 1: See Average Coverage Percentage
First of all, pay attention to Average Coverage percentage. Rather, what percentage of the potential economies of 80-90% RIs deployment are economically desirable? More coverage means more tires are used on the RIs and therefore more savings are achieved. If composite coverage is considerably below these figures, that means potentially a lot of On-Demand workloads which are more expensive are being deployed.
Step 2: Identify Reserved Instances without any Usage
Next, you need to search for the RIs which have remained underutilized. In such cases, the instances are at 0% coverage, and are fully “on-demand”. Consequently unnecessary costs are incurred because on-demand costs can be 50% (sometimes more) expensive than reserved. Such unutilized RIs can be adjusted to correspond with your current workload and made use of or alternatively, they can be sold at the RI Marketplace.
Optimizing RI Coverage
To achieve optimal Reserved Instance (RI) coverage, there is a need for a strategy and planning which in turn greatly improves cost savings and performance of your cloud infrastructure. Luckily there are several RI covering strategies, complete with the examples and a dash of humor, which will elevate your RI coverage:
Importance of Matching RI Purchases to Usage Patterns
Interfaces are critical for the users intending to utilize the benefits of RIs optimally as appropriate purchases should match actual usage. For example, while most of your historical records show steady usage, there is a notable peak during the year end holiday season – probably your online store just gets overloaded with people wanting to buy things: during the holiday, you want to reserve extra silver instances. It is like ordering pizzas for a party, everybody has decorations and nobody wants to run out of it.
Strategies to Improve RI Coverage
Analyze Historical Usage Data: Only as far back as their historical usage data allows them—Which in this case is very useful so they will have to delve back deeper seeking more sisyphean challenge analysis. Try to use the AWS Cost Explorer to understand different patterns, for example, an increase in usage during that time of the year, which is due to world activities such as the World Cup, FIFA and Olympics. Since the figures usually take such a form, adjust your RI purchase accordingly as that project which always seems to come up during summer and no one wants the issues associated with it but everyone has to undertake it somehow seems unavoidable.
Forecast Future Usage Needs: Whenever a provider begins thinking about usage expectations, which by the way, is an extremely important step in planning, they always have to deal with historical data, and look ahead - as for example, how many packets of chips do you need for watching the third Star Wars on the tenth viewing. Encourage stakeholders to provide information on how the workload will shift, for instance, if some plans to roll out a new enhancement which would need more manpower. For example, consider this as being able to receive your visitors, in this case, your in-laws, without having to panic that you will run out of snacks, but at the same time you want to ensure that no one overbought snacks!
Buy RI Purchasing Based on Usage Thresholds: Buying Reserved Instances (RI) makes this challenge easy! Set your usage thresholds in place, and you're done, much like a self-watering plant; now, your resources will be met without continuous checks on them. That is you are better placed for more important tasks, such as recalling where you placed that peculiar vegetable you unwittingly bought.
Monitor RI Coverage and Utilization Regularly: Ensure that you actively manage RI utilization by using AWS Cost explorer. Do not forget to review on a periodic basis – it would take you days to self-report yourself through the fridge and noticing that there are some items left unattended. If you can find a few RIs only sitting there cause you did not take them out for use, then it is time for a strategy check. .
Or, if all of these steps seem like a lot of work, you can alternatively try Pump for free. We help customers save on their AWS for no cost, and automate buying / selling RIs.
Conclusion
In adopting the policy to improve and refine the cost analysis of the EC2 RI Coverage Report readers can lower costs and hence maximize their benefits from the AWS reserved instances. You will be able to look at different dimensions of the existing metrics, and find places where changes will bring more money to your company, and include additional purchasing and control over the selected strategies.
Remember to improve your RI Coverage Report on a regular basis, if possible, and modify your RI strategy in accordance with the growing workload. Don't delay in optimizing your RI coverage, and start saving your business the substantial cost.
Join Pump for Free
If you found this post interesting, consider checking out Pump, which can save you up to 60% off AWS for early-stage startups, and it’s completely free (yes, that's right!). Pump has tailor-made solutions to take you in control of your Amazon EC2 cost and all other AWS spend in an effective way. So, are you ready to take charge of cloud expenses and maximize the most from your investment in AWS? Learn more here.
