Maximise your Cloud Cost Savings with these Tried-and-Tested Strategies that helped us achieve a 42% reduction in the Google Cloud?Spend

Nalina Madhav c , Swathy Bharathan , Tushar Bhattacharya

Say goodbye to unnecessary resource usage and hello to cost savings and efficiency..!

It’s no secret that the Cloud usage can quickly add up, and without proper optimisation, cloud bills can spiral out of control. Managing cloud costs can be a challenging task, especially as usage grows and expands. It’s crucial to take control of your cloud expenses and find ways to reduce costs without sacrificing performance.

At airasia Super App, Flights team faced similar challenges with our GCP expenses as we began to expand the business to support more products and features to sell airasia + 700 airlines. We found that our cloud bills were significantly higher than we anticipated. But we didn’t let that deter us. Instead we dug in and started exploring different cost-saving techniques.

In this blog post, we’ll explore some of those tried-and-tested strategies that can help you save money on your GCP usage. So whether you are new to GCP or a seasoned pro, if you’re ready to take control of your cloud spending and to start optimising your GCP usage, read on to discover how you can save money on your GCP bills.

6 simple steps to kickstart your cost optimisation journey

Our Cost Optimisation journey commenced with a thorough analysis of the billing report in order to pinpoint the primary cost drivers. Google’s GCP Billing Report provides a comprehensive breakdown of costs by the SKUs that served as substantial contributors to the overall GCP bills. With this, we were able to swiftly and accurately isolate the significant cost drivers, enabling us to make informed decisions and take appropriate measures to optimise our expenditure.

Yes, you read the Pie Chart right!?:)?

Our Logging was one of the culprits behind the massive expenses that had us gasping in disbelief. The seemingly harmless activity of logging turned out to be the real-life monster that devoured our budget, and we’re still reeling from the shock.

Based on the above chart, we decided to segment the optimisation strategies into two key areas

1. Cloud Logging optimisation
2. Infrastructure cost optimisation

Cloud Logging optimisation

Logging is an essential part of the Software development, but it doesn’t have to come at a high?cost.

By using the below mentioned simple but cost effective measures, we were able to unlock significant cost savings while maintaining the level of visibility and insights we needed to keep our systems running smoothly.

Fine tuning the retention policies and leveraging exclusion filters allowed us to strike a balance between operational efficiency and cost management, enabling us to keep our cloud logging expenses under control. With this, we successfully managed to slash our logging costs by a whopping 70%.?

The below graph shows the cost trend before and after optimisation. The results speak for themselves — our wallets are feeling much fuller, and we can now channel those savings towards other areas of our business.?

Cutting the Clutter: How Retention Policies and Exclusion Filters cut our Logging Costs by?70%

In the next section, we’ll look at some of the strategies we devised and implemented to contain infrastructure cost.

Infrastructure cost optimisation

CUD Plan: One of our easiest choices was to agree to a committed usage discount plan. As our services are deployed in a GKE Autopilot environment, we were entitled to purchase spend-based committed use discounts.

Committing to Savings: CUD plans dropped our GKE Costs by at least?20%

Usage of Spot VMs: We took advantage of the Spot instances to achieve significant cost savings on our non-critical workloads.

Spot On: Leveraging Spot VMs slashed our Daily Costs by?50%

On Demand Environment: We optimised non-production environment usage by bringing the pre-production environment online only when needed while scheduling regular shutdowns for our development environments during off-hours, such as nights and weekends.

Dynamic Resource Provisioning: We implemented automation to streamline our pre-sale cluster warm up and scale down process, reducing manual labor and minimising the costs for the non-peak sales hour.

Container Image pruning: We integrated a CI job to get rid of old container images during every service deployment to minimise container image storage cost.

Feeling a little lost in all the optimisation jargons? No worries, I’ve got your back with the ultimate cheat sheet to unlocking your savings!?:D?

Cheatsheet to Unlocking your GCP?savings

Stay tuned, because the optimisation journey is far from over! In our upcoming blogs, we’ll deep dive to further understand the strategies mentioned in this blog post.