In the last newsletter I was talking about the new managed Kubernetes from Google. And recently, I read this one: "No one wants to manage Kubernetes anymore". And I cannot agree more. It seems to me there is a community of geeks who swear by self-managed kubernetes. But why would you do this? That's awfully complicated, time consuming and risky.
Then, this question: "Are Containers Still Relevant?". I think so but definitely not self-managed. Read those articles and give me your thoughts.

The Well-Architected Framework and the Serverless Lens should be in your favorites already, and if you didn't read them, please take the time, as it provides mountains of recommendations. And if you need a summary, you can read this article.

If you want other best practices, you can look at those patterns, provided by the AWS Developer Advocates. Each one is provided with a SAM template so you can easily implement it. Another great source of information is the new Lambda Operator Guide. It covers concepts and best practices for designing, building, measuring, troubleshooting and optimizing Lambda functions. Another link to put in your favorites.

You can choose to follow these 28 experts or you can just follow me :) Just kidding, those men and women are just incredible. Most of the great content I published in this newsletter come from these people. However, someone important is missing: James Beswick. He's delivering lots of content, many AWS blog posts, he's the author of the Operator Guide mentioned above, and much more. Sheen Brisals would probably also deserve a place in this list.

Talking about James Beswick, here you will find a great series of articles on the AWS Blog, dedicated to monitoring and observability for Lambda-based applications: CloudWatch, Log Insights, custom metrics, X-Ray, ...

Recently, AWS announced a new feature that combines S3 and Lambda: S3 Object Lambda. Using a new type of access point, you can have a Lambda function acting as a proxy between your S3 bucket and the client requesting some objects. I wrote an article to explain how to use this feature to anonymise data.

Secrets Manager is the way to go when you need to store some secrets, for serverless applications but not only. This article guides you through the implementation in a Lambda function using javascript. If you use python or java, I would rather recommend the lambda-powertools libraries, but since it's not yet available in js, great example.

S3 notification to Lambda is one of the first triggers provided when Lambda was first introduced in 2014. But there is a limitation: you cannot define more than one event rule for the same event type (ex: ObjectCreated). This article provides several solutions to overcome this limitation.

If you didn't watch the video on Lambda function isolation from last re:Invent, then I strongly recommend to watch this one. Marc Brooker is really good at explaining Lambda under the hoods. And great chat with Marcia Villalba.

In this article, Lucas de Mitri gives interesting feedback on some challenges you can face with Lambda in production: database connections, scalability of downstream components, infra as code.

Once you know what should be tested in a serverless application, you can have a look at localstack and this article to test your serverless workload locally. I've been trying it in the past but was not convinced: complicated to set up, randomly fails, and most important it does not always match the real behaviour. I generally recommend to perform integration tests on the cloud, and keep local for unit tests. But give a try, maybe it will fit your needs.