CloudForecast Streamlines Operations with Judoscale
In the world of AWS cost management, CloudForecast helps businesses keep their cloud expenses in check. But even experts sometimes need a hand with their own operations. I sat down with Francois Lagier, CTO of CloudForecast, to discuss how Judoscale helped them focus on what really matters—building their product.
CloudForecast’s Tech Stack
CloudForecast runs a hybrid infrastructure. Their main application is a Ruby on Rails app hosted on Heroku, while their more intensive background jobs run on Amazon ECS. Francois explains the reasoning behind this setup:
We were running into some memory issues on Heroku. For the price, ECS in some ways is… we are getting better performance for the cost. Also, our entire team is well-versed in AWS and moving to ECS wasn’t that complicated for us.
The Challenge: Efficient Scaling
CloudForecast’s background jobs, particularly those handling daily email reports, experience predictable spikes in activity. Before Judoscale, they managed this with CloudWatch alarms set to scale up resources 15 minutes before their usual 2 AM GMT peak.
However, this solution wasn’t perfect. Francois noted:
We knew we had a spike, knew we needed more resources at that time, but every now and then during the day, we will also have a spike.
These unexpected spikes during the day meant that sometimes, important information for clients could be delayed in processing queues.
Enter Judoscale: Automating the Process
For Francois and his team, the goal wasn’t just about reducing costs - it was about simplifying operations and focusing on their core product. They turned to Judoscale for a solution that would work seamlessly with their current stack.
I believe that every company needs to be good at specific things. Figuring out auto-scaling for us is not a priority. CloudForecast is not a better product if my team or myself are [building] our own thing.
The integration process was straightforward. Francois estimates it took less than an hour, with most of that time spent on deployment rather than actual setup.
The Results: Set It and Forget It
Since implementing Judoscale, CloudForecast has experienced several benefits:
- Reduced manual intervention: No more alerts about overloaded queues.
- Improved processing times: Spikes in activity are handled automatically.
- Focus on core business: The team can concentrate on improving CloudForecast.
Francois appreciates the hands-off nature of the solution:
We haven’t thought about it since we set it up like six months ago. It’s one of the things that just runs, we don’t really have any issues with it, which is nice.
While cost savings weren’t the primary goal, CloudForecast has seen some reduction in their AWS expenses. More importantly, they’ve optimized their resource usage, aligning with their own mission of efficient cloud management.
The Verdict
When asked to summarize his experience with Judoscale, Francois was clear:
First of all, it’s working. The integration is easy. It does exactly what it’s supposed to do. We never run into an issue… It’s a no-brainer. It’s one of those products where if you know you’re running a lot of things on your local Sidekiq or ECS or whatever, it’s not costing you more money, it’s actually saving you money. It’s one of those products that just makes sense.
For CloudForecast, Judoscale has proven to be a valuable tool in their tech stack. It’s allowed them to automate a crucial aspect of their operations, freeing up time and mental energy to focus on what they do best - helping their clients manage AWS costs.
If you’re looking to streamline your ECS scaling operations and focus more on your core product, Judoscale might just be the solution you need.