Google Cloud drops new GCP developer cheat sheet

This has been a big month for Google Cloud Platform, including new Cloud Deploy and Certificate Manager products, autoscaling for Bigtable, Assured Workloads for EU, a new generation of Cloud Functions, and an updated Google Cloud Developer Cheat Sheet and GCP Architecture Diagramming Tool! Let's dig in!

Want all this month's Google Cloud news in video form? Check out our the most recent episode of GCP This Month.

Your keys to a better career

Get started with ACG today to transform your career with courses and real hands-on labs in AWS, Microsoft Azure, Google Cloud, and beyond.

The Google Developer Relations team has been hard at work, and they have made a massive update to the Google Cloud Developer Cheat Sheet. If you’ve taken some of my Google Cloud training, then you’re probably already familiar with this amazing tool. And if you haven’t, then today’s your lucky day!

Because this thing is awesome!

The cheat sheet is also known as “Google Cloud products in 4 words or less,” and that is a part of its beauty. These judiciously chosen words give you a good feel for what a particular GCP product does—actually, what every product does, because it’s comprehensive!

The printable poster version is still my favorite, but a key benefit of the new interactive version is that every product card links you straight to the documentation so you can dig into that if you need more than just the four words. And the list view—selectable at the top—has the same comfortable at-a-glance feel of the poster.

And not only have we gotten a massive update to this immensely valuable tool, it looks like we’re going to be getting a lot faster updates to it in the future, too! Priyanka Vergadia regularly puts out great content about Google Cloud, and since she has now taken over maintenance of this tool, its future is looking bright!

But wait, there's more!

They’ve also just released a new diagramming tool for GCP architectures! By itself, this doesn’t seem so exciting. But the coolness is in the integrations!

First, the components for your diagram are connected with the cheat sheet to help with both picking and understanding them. Next, you can bootstrap your diagrams from a library of reference architectures! Not only that, some of these architecture examples are set up so you can very easily deploy them into your project to play around with them!

Nice, eh? I thought you’d like that. And here’s a whole slew of other GCP quick bites for your enjoyment.

If you’re using Apache Spark to squeeze value from your data, I’m sure it’s not because you just really love managing those Spark clusters, right? Well, you’ll be very happy to hear that Google has just released their Dataproc Serverless for Spark offering for general availability! Now you can let Google manage the clusters and tune the infrastructure, and you can focus on coding instead.

As a bonus, there’s now also a private preview of Spark through BigQuery, making it even easier to get at that Spark-y goodness, right close to your data.

Now, speaking of managing clusters, Bigtable has also gotten a boost. Autoscaling for Bigtable is now generally available. There’s no point in paying for an overprovisioned Bigtable cluster when your system doesn’t need all that capacity, so this is a welcome change to better match your costs to your actual usage.

And although Google’s blog post uses the rather fun word “diurnal” no less than six times, I’m sure you could just as easily save the 40% they mention if your workload happens to be “nocturnal” instead.

OK… This announcement is about the new Compute Optimized C2D instance family using 3rd Gen AMD EPYC Milan processors. You can beef these up with as many as 112 vCPUs and 896 GB of memory. Of course, even though these are compute-optimized CPUs, you can still choose your most appropriate ratio of processors to memory, as usual, in Compute Engine.

In case you’re unfamiliar with it, Assured Workloads is Google’s response to Amazon’s GovCloud—but instead of being a separate cloud region that meets compliance, Google has a system for workloads to meet compliance in any GCP region within that part of the world. And now Google’s offering of Assured Workloads for the European Union is officially GA.

While this includes data residency, data encryption, and Google personnel, they are still in public preview with the Sovereignty Controls integration with Cloud External Key Manager that lets you keep your data encryption keys outside of GCP.

If you’ve been following this series for a while—or hey, maybe even if you haven’t—then you probably know that I love all things serverless! And if you do too, then you’ll be happy to hear that Google has just released into public preview their second generation of Cloud Functions!

Now, since this new version is built on Cloud Run under the hood, this is really blurring the lines between the serverless compute options. So, we still have the same pay-for-use compute as the first version of Cloud Functions, but now we can also allow for simultaneous processing of multiple requests/events, just like Cloud Run!

Of course, before you turn this on, make sure your code can handle concurrent requests! But this could save you a bunch of money if your event handling isn’t always pegging the CPU. Also, this second generation lets you:

• run your functions for longer and on beefier instances
• pin instances to reduce cold starts
• split traffic between multiple versions

These new Cloud Functions, they’re grrrrrrrrrreat!

Now, it’s no secret that all websites had all better be using HTTPS these days. I mean, not only will normal HTTP connections be marked as “Not Secure” in the Chrome address bar, but insecure sites also get buried pages deep in Google searches. So it’s great news that Google now has a new serverless Certificate Manager product in public preview.

Certificate Manager makes it super-easy for anyone to get and use TLS certificates on the external HTTPS load balancers fronting their GCP-based workloads. You can use 100 certificates for free each month, and it’s pay-per-cert after that.

And because these get managed by Google’s infrastructure, your clients can quickly negotiate secure connections with Google’s edge nodes. Super-low latency! Oh, and this product also supports wildcard certificates. Pretty wild, eh? OK, I know that joke was terrible. I’m sorry.

Another new product is Cloud Deploy, also now GA! As Google notes in their announcement, “​​The operational cost of Kubernetes continuous delivery can be very high.” Well, Cloud Deploy aims to take some of that pain away with this managed and opinionated continuous delivery service for GKE.

If you’re using GKE, then you might find it well worth the extra fifteen bucks per monthly delivery pipeline beyond the first free one. Cloud Deploy is built on Skaffold and includes both easy integrations to GCP management and also many built-in best practices.

Finally, I want to let you know about some upcoming events. Coming up on April 6 is the Data Cloud Summit. This will be an online event all about “AI, machine learning, analytics, databases, and more.”

And even better, we now have a date for Google Cloud Next ‘22! Cloud Next is Google’s annual conference packed with all the most exciting GCP announcements, so make sure to block off October 11-13. And if you’re hoping against hope, like I am, that this might wind up being an in-person event, then be sure to reserve some extra time for travel to and from San Francisco. Fingers crossed, and I hope to see you there!

Subscribe to A Cloud Guru on YouTube for weekly news and updates. You can also like us on Facebook, follow us on Twitter, or join the conversation on Discord!

Stay safe, take care of those around you, and keep being awesome, cloud gurus!