Five Overlooked Cloud Bill Red Flags and Tips to Handle Them

Uncover hidden cloud billing issues with these tips: policies, API control, logging, and questioning bills.

March 18, 2024

Cloud cost optimization requires understanding your bill. Sam Clark, technical account manager at DoiT International, identifies five often overlooked reasons that could be causing your cloud bill to spike and ways to address the issues. 

Have you ever really dissected your cloud bill? When optimizing cloud costs, we often focus on major spikes only, failing to heed the more subtle signals hiding amidst the numbers. Those less obvious red flags indicate something is amiss, potentially causing an organization to overspend or operate less efficiently.  

Users very often overlook billing issues with their cloud spend. Here are five that often fly below the radar and tips for resolving them.  

1. Implement policies

If your cloud storage costs steadily rise, appropriate object lifecycle policies are not in place. These can automate transitioning data to various storage tiers while deleting data via predefined rules. This aligns storage costs with the value and accessibility of the data, preventing a business from overpaying for information that doesn’t require immediate access or is simply no longer needed.

With lifecycle policies, you avoid a build-up of data, an ever-increasing log store, and possibly an excessive volume of snapshots. This also drives up storage costs, particularly if older or less frequently accessed data is held in high-cost tiers. Typically, transitioning or expiring objects after 30-90 days is fine. Still, when costs are trending higher, it’s time to dig deeper into this area.

2. Follow the CloudTrail 

If you’re paying for AWS CloudTrail, you already have a chance to cut costs on your cloud bill. While the initial trial in a region is free, the only one should be at the AWS Organization level because these are automatically created in all member accounts. Remember, new trails are additional to existing ones, and if you already have existing separate trails, you can eliminate those to lower costs.

What’s more, with a trial at the organizational level, you can uniformly apply and enforce an event logging strategy because configuration propagates to all accounts. That said, be sure to verify that the configuration of your Organization trail matches how you want all trails to be.

3. Cross-check your decisions

Failing to cross-check technology and design decisions can snowball into major cloud bill issues and create performance problems. An example would be when a team buys a commitment discount without considering broader organizational strategy or the needs of other groups. In this case, the company is stuck with virtual machines if someone purchases a multi-year reserved instance – but there’s a strategic push towards serverless in the next few months. 

Simply put, cloud-related decisions made in isolation can raise costs and reduce efficiency. So, check internally to ensure decisions are made with the good of all in mind. This could mean cross-checking cloud infrastructure decisions against engineering strategy and vision.

4. Control high API costs 

Third-party services such as New Relic and Datadog can scan cloud accounts regularly to provide the latest usage information. Unfortunately, many organizations aren’t aware that they pay for any API requests these services require. These requests are reflected in the GetMetricDataOpens a new window API SKU, and if you’re not careful, you could shell out a lot of money for API calls coming from the third-party software. 

This makes monitoring the frequency of API calls and the metrics and data being scanned critical. For example, you could have a development account with many resources eating up a lot of money on Cloudwatch, courtesy of an API call every minute. Is this frequency and granularity of data needed? 

To bring this under control, you must ask the third-party providers to adjust the frequency and number of metrics being pulled for specific accounts and projects. 

5. Cut back on logging 

Logging is required for monitoring and troubleshooting. However, excessive logging can seriously add to your cloud bill. Like the API calls mentioned previously, you must ascertain whether the frequency and metrics you’re gathering from logging are right for a particular use case. For instance, if you’re feeding a dashboard with log data, you likely don’t need per-second updates, especially when once every five minutes would work fine.

Generally, logging should not exceed 20% of your cloud bill. If you’re over that mark, it’s a clear indicator these costs need a closer look. With this in mind, finding out how various teams utilize your logs can help you determine a better frequency and arrive at more useful logging metrics.

Further, stay abreast of logging in non-production environments because they usually don’t generate money. You likely won’t need the same verbosity or history required in production accounts. And suppose something does break in a non-production environment. In that case, you can simply switch logs on and off, a far cry from a production environment in which additional information is needed to figure out why something has broken.

See More: Building Without APIs Comes with a Cost

Question Your Bill

These are just some of the top red flags in a seemingly endless list. With this in mind, consider your cloud spend and dig into the details. Don’t assume your bill is what it is; challenge issues, ask questions, and repeat this process. 

For instance, if S3 costs are rising, find out what’s causing it. Follow up by asking which SKUs are responsible for the increase. If you find out it was due to data transfer, ask your team if it was anticipated. The costs could result from a valid reason, but you’ll never know unless you ask.  

Doing so can help build a cost optimization culture across your company, where everyone recognizes their impact on the bill and can take appropriate action. Shared responsibility is required to control cloud costs effectively, and it’s a process that works best when approached by the full team.

How do you manage your cloud costs? Let us know on FacebookOpens a new window , XOpens a new window , and LinkedInOpens a new window . We’d love to hear from you!

Image Source: Shutterstock

MORE ON CLOUD SPEND

 

Sam Clark
Sam Clark

Senior Technical Account Manager, DoiT International

Sam Clark is a senior technical account manager at DoiT International. He possesses more than 25 years of experience in the technology sector, with a dedicated focus on cloud expertise spanning the last decade. Specializing in optimizing cloud operations and cost efficiency, he champions the principles of FinOps via webinars, podcasts and blog posts.
Take me to Community
Do you still have questions? Head over to the Spiceworks Community to find answers.