Preview zone: How Google Cloud CRE helps SaaS companies prevent unanticipated failures

At Google, our Customer Reliability Engineering (CRE) teams work with customers to help implement Site Reliability Engineering (SRE) practices to continually attain their reliability goals. This work often includes defining objectives and implementing operational best practices like blameless postmortems or analyzing error budget spend.

Following CRE practices is especially important when changes are made in the customer's product. But what about when changes are released within Google Cloud Platform (GCP), where the product runs? We’ve heard that you want to test your products against future GCP releases to ensure reliability and performance when the underlying cloud service changes. We are happy to announce that preview zones are now available to let you test your own production code against future releases of GCP.

We’ve been working recently with many of our SaaS company partners and we’re happy to announce that we’ve expanded our CRE for SaaS program to address these needs. You can see how it works here:

With this expansion, our SaaS partners who have enrolled in our CRE for SaaS program now have an option to run a copy of their production applications in the preview zone. This lets partners detect unanticipated failures of applications running on future releases of GCP services. We put a number of unreleased “Day 0 binaries,” our soon-to-be-released code, in this zone. Then partners can test their production applications against that code. This way, we can anticipate and avoid previously unknown failure modes before users do—giving both us and our partners a chance to investigate the pending changes and address them.

BrightInsight (a Flex company), this year’s winner of the Google Cloud Healthcare Partner award, has been using the preview zone, and finds it helpful both in preventing unanticipated failures as well as supporting regulatory compliance requirements within the healthcare industry.

To use the preview zone, you’ll need to have defined your SLOs so that Google can integrate them with additional test frameworks. If you don’t have SLOs defined, we’ve built SLO Guide, a new tool to help you discover what you should measure based on common architectures and critical user journeys. It will help you quickly create SLOs that measure what your users actually care about. You can request access to the tool here. Finally, if you’re not a Google Cloud SaaS partner yet, kick off the process here.