.png)
Lars Francke,
CTO and Co-Founder
CTO and Co-Founder
"A single cluster running in Azure just to run our integration tests used to cost us between 6... 7... 8,000 euros a month. This is an awesome alternative. You can get an OpenShift cluster in seconds because there are hot clusters running all the time. We can integrate them into our github actions, get a cluster, pay just for the seconds we use it and then drop it again."
Company Size:
20
Industry:
Data Management
Challenge
Stackable is an open-source platform that helps companies manage large amounts of data. It can be used to create data warehouses, data lakehouses, event streaming, machine learning, and data meshes. Stackable’s platform is Kubernetes native, and can be used both on-prem or in the cloud.
For Stackable’s on-premises product, they understand how important it is for their customers to have smooth installation and upgrade experiences. With customers on many different distributions - like OpenShift, AWS, GKE, and more - they decided to invest in building a tool that would help them to test their application against the different distributions. This way, they could be certain that what they shipped to their customers could be successfully installed regardless of Kubernetes version or distribution type.
However, Stackable quickly found it difficult to keep up with the constant changes in each distribution. For example, when AWS released a new version of their managed Kubernetes, they removed a storage class and it completely broke Stackable’s testing system. “We need to consistently run those tests against as many Kubernetes flavors or variants as we can find to make sure we catch those things before our customers catch them,” says Sönke Liebau, Co-Founder and CPO. Although their homegrown solution was becoming too much to manage, both financially and in development hours, it wasn’t an option to eliminate testing their application across different software distributions.
They began their search for another solution, and discovered Replicated’s Compatibility Matrix.
For Stackable’s on-premises product, they understand how important it is for their customers to have smooth installation and upgrade experiences. With customers on many different distributions - like OpenShift, AWS, GKE, and more - they decided to invest in building a tool that would help them to test their application against the different distributions. This way, they could be certain that what they shipped to their customers could be successfully installed regardless of Kubernetes version or distribution type.
However, Stackable quickly found it difficult to keep up with the constant changes in each distribution. For example, when AWS released a new version of their managed Kubernetes, they removed a storage class and it completely broke Stackable’s testing system. “We need to consistently run those tests against as many Kubernetes flavors or variants as we can find to make sure we catch those things before our customers catch them,” says Sönke Liebau, Co-Founder and CPO. Although their homegrown solution was becoming too much to manage, both financially and in development hours, it wasn’t an option to eliminate testing their application across different software distributions.
They began their search for another solution, and discovered Replicated’s Compatibility Matrix.
“[Our homegrown solution] worked for us, but it was difficult to upkeep… whenever we wanted to add a new provider there was a significant amount of work to write Terraform scripts that worked and integrate that with our stuff. Compatibility Matrix just makes that much nicer for us.”
.png)
Sönke Liebau,
CPO and Co-founder
CPO and Co-founder
Solution
Replicated’s Compatibility Matrix creates customer-representative environments that companies can use to test their application against by combining different versions of Kubernetes, distributions (such as OpenShift, AKS, GKE), and instance types. By utilizing Compatibility Matrix, companies can provision the infrastructure to build confidence their software will deploy successfully in current customer environments, as well as recreate customer-representative environments when support escalations do occur.
When Stackable switched over to Compatibility Matrix from their homegrown solution, their team immediately experienced a positive impact. They set up the same tests within Compatibility Matrix that were once in their homegrown solution, and in Compatibility Matrix found they were 2X faster. Tests that previously took 40 minutes to run now only took 20, giving back time to their development team.
In addition, they were able to test across new distributions that were previously too difficult and time consuming to build in their own testing solution. Most notably, their team now had the ability to test how their application would be installed by customers using OpenShift - notoriously one of the most difficult distributions to install into and heavily used by some of the largest finance companies, utility companies and federal agencies in the world.
One of the biggest benefits of Compatibility Matrix is that you only pay for the time you spend actively testing your application on the clusters. By moving to Compatibility Matrix, Stackable was able to save thousands of dollars a month. Previously, they were spending up to $7,000 per month just running Azure clusters to test against. By moving over to Compatibility Matrix, that spend dropped down to just a few hundred dollars at most.
By moving to Compatibility Matrix from their homegrown testing solution, Stackable was able to easily test across more distributions, and greatly increase confidence that their end customer installs would be successful. They were able to take the management and maintenance of distribution testing off their plate, and give themselves time back to focus on building their core product while simultaneously saving thousands of dollars a month.
When Stackable switched over to Compatibility Matrix from their homegrown solution, their team immediately experienced a positive impact. They set up the same tests within Compatibility Matrix that were once in their homegrown solution, and in Compatibility Matrix found they were 2X faster. Tests that previously took 40 minutes to run now only took 20, giving back time to their development team.
In addition, they were able to test across new distributions that were previously too difficult and time consuming to build in their own testing solution. Most notably, their team now had the ability to test how their application would be installed by customers using OpenShift - notoriously one of the most difficult distributions to install into and heavily used by some of the largest finance companies, utility companies and federal agencies in the world.
One of the biggest benefits of Compatibility Matrix is that you only pay for the time you spend actively testing your application on the clusters. By moving to Compatibility Matrix, Stackable was able to save thousands of dollars a month. Previously, they were spending up to $7,000 per month just running Azure clusters to test against. By moving over to Compatibility Matrix, that spend dropped down to just a few hundred dollars at most.
By moving to Compatibility Matrix from their homegrown testing solution, Stackable was able to easily test across more distributions, and greatly increase confidence that their end customer installs would be successful. They were able to take the management and maintenance of distribution testing off their plate, and give themselves time back to focus on building their core product while simultaneously saving thousands of dollars a month.
Using Compatibility Matrix has reduced test run times by half, from 40 to 20 minutes.
“Compatibility Matrix offers more distros and versions than we had before - we didn’t have AWS and OpenShift, which was a huge game changer for us. Plus, Compatibility Matrix lets you pick your specific version of these distributions which we couldn’t do before.”
Sönke Liebau,
CPO and Co-founder
CPO and Co-founder
%201.png)