The overwhelmed person’s guide to Google Cloud: week of Nov 13

The content in this blog post was originally published last week as a members-only email to the Google Cloud Innovators community. To get this content directly in your inbox (not to mention lots of other benefits), sign up to be an Innovator today.

New and shiny

Three new things to know this week

• Vector inspector. The artist formerly known as Vertex AI Matching Engine is now simply called Vector Search, and we just released a bunch of new features to make it easier to explore, filter, and get started doing cool things with vector embeddings. To the vector be the spoils!
• Remote GOAT. Remote and virtual repositories are now generally available in Artifact Registry. Remote repos can locally cache artifacts, insulating your build process from downtime; virtual repos let you manage internal and external endpoints under a single URL. Greatest artifact registry of all time? I didn’t say it.
• Enter-prized. GKE Enterprise is now generally available. To some extent you can think of this as “all the things you know and love about GKE, now wrapped up in a package that works for large organizations”, but there are some legitimately cool new things in Enterprise as well. As I mentioned in the developer keynote at Next, one of my favorites is Advanced Vulnerability Insights, which does real-time vulnerability scanning on your containers all the way down to the language-package level.

The hot seat

Behind the scenes of Google Cloud with the people who build it

SLO and steady: SRE turns 20

Google Site Reliability Engineering (SRE) just celebrated its 20th year of reimagining software delivery and operations at scale. Jennifer Petoff, longtime director of SRE Education at Google and co-editor of the SRE Book, joined the newsletter to reflect on the legacy (and future) of the SRE role.

Forrest: Jennifer, thanks for all your work helping teach the world the principles of SRE. What are a couple of common qualities that show up in the great SREs you’ve worked with over the years?

Jennifer Petoff: SREs are fantastic problem solvers and they don’t give up easily. I always think about the old TV show MacGyver. Week after week, MacGyver used his scrappiness and know-how to resolve critical situations. I think SREs are great at “MacGyvering” solutions to tough production problems. That said, SREs are never satisfied with point fixes, they tirelessly develop fixes to entire classes of problems to make sure they don’t happen again.

I’ve also found that SREs are some of the most collaborative and supportive people I’ve ever worked with. I think it’s a necessary trait for the job. Running production services requires SREs to work across teams, to work across functions in order to deliver our best to our users.

What’s a misconception people commonly have about SRE, even after 20 years?

Some people and organizations think that SRE only works if you are Google or operating at Google scale. While there are many things associated with an SRE practice that you might not want to do if you are a 50 person start-up (e.g., having an entirely separate SRE organization), the foundational principles always apply and can help you drive continuous improvement to your production services and in effect elevate your customers’ experience.

You co-edited the original SRE Book, which has been influential in organizations far beyond the tech industry. Why do you think this book has had such an impact?

I think the book really resonates because the foundational principles (e.g., SLOs, error budgets, blameless postmortem culture) are so rational and applicable to a wide range of situations.

For example, SLOs help different members of an organization representing different parts of the product life cycle speak the same language and agree on what “reliable enough” means before there is a crisis. This helps take the emotion out of decision-making to help organizations run more smoothly.

I think the book also resonates because it advocates approaches on how to work smarter, not harder. For example, SRE really focuses on developing engineering solutions that enable scale. Let’s not feed the machines with human toil!

You can hear more reflections from Jennifer and many longtime Google SREs in the 20th-anniversary video series here.

Watch this

Here’s how you can use Duet AI to speed up creating a weather app with Python and Flask.

Community cuts

Every week I round up some of my favorite links from builders around the Google Cloud-iverse. Want to see your blog or video in the next issue? Drop me a line!

• RAG (Retrieval-Augmented Generation) is the latest generative AI buzzword, and Christoph Bussler shows how to turn it into reality using AlloyDB AI.
• Champion Innovator Onkar Naik digs deeper into slinging Terraform with the new Infrastructure Manager service.
• The Kubernetes Gateway API reached v1.0 this past week, a major milestone for the open-source community. Mathieu Benoit has a helpful exploration of how the Gateway API fits together with Istio and the Anthos Service Mesh.

Learn and grow

Four ways to build your cloud muscles this week

• Learn and burn. “No cost November” is a thing, I guess? Check out November’s on-demand Google Cloud trainings, including a healthy helping of generative AI stuff.
• Cache rules everything around me. I always love these deep dives from the Google Cloud product engineering teams. Here’s how they souped up Redis scaling for the new MemoryStore for Redis Cluster service.
• Pool party. You want to optimize compute costs in GKE? Then you want to use node pools with priority-based scheduling, and this blog shows you how.
• Bet on Duet. I happen to know we have a lot of Google Workspace gurus reading this newsletter, and those people should all join us on November 30 to learn how Duet AI is making Workspace more awesome.

One more thing

Become an Innovator to stay up-to-date on the latest news, product updates, events, and learning opportunities with Google Cloud.