Published in


Recap: Re:invent 2018: It was all serverless this time!

It has been 2 weeks after Re:Invent. Last week, as I recovered from the Las Vegas-to-Turkey trip jet lag, I took some time to evaluate what we learned from the event, made some adjustments to the Thundra roadmap, and got back into the daily work routine. Now, I finally have some time to share my thoughts from the event with this recap blog.

First, the AWS team does a great job putting on the conference. Previously, I followed Re:Invent via live streams but this was my first in Vegas and I really admired the quality of work by the team to manage, educate, and entertain tens of thousands of attendees.

I was often on duty at our Thundra booth, giving demos to visitors curious about AWS Lambda, Thundra, or simply just curious. Traffic was non-stop, giving me the opportunity to validate assumptions and clarify some open questions I had gathered before coming to Re:Invent. I want to thank the +1300 visitors who stopped by our booth and shared their experiences and needs regarding serverless, monitoring or both.

What did I learn from our discussions?

Many people are in the early stage of serverless adoption, which means there were lots of questions about AWS Lambda and serverless. Importantly, there were also many questions about how to monitor serverless applications and why monitoring is important when just starting out. Many many people didn’t realize that there are tools beyond CloudWatch that can be used to monitor AWS Lambda functions. In fact, development times will be greatly reduced when they expand their monitoring to include tools that provide a much greater context, and greater detail. With a comprehensive observability solution, AWS Lambda developers can fix issues faster, optimize their resource consumption (and control costs), and quickly discover bottlenecks in their serverless environment — all without guessing or digging through the details of each log.

Once we showed how you could easily accomplish this with Thundra *and* explained that Thundra was free to use, I was glad to see that visitors responded with delight and offered up additional ideas of their own. I plugged in some of the most popular feedback into our roadmap and am looking forward to seeing these ideas come to life in future versions of Thundra.

Several Important Serverless Announcements

Both Thundra and AWS made important serverless-related announcements during re:Invent week, including one major announcement that Thundra and AWS made together.

Among the many announcements, the following were the ones that I believe all of us interested in serverless should take note of:

  • AWS announced the Python 3.7, Ruby, C++, and Rust support for AWS Lambda. I believe this shows their dedication to expanding the serverless community to include more and more developers. With this additional support, AWS Lambda puts themselves way ahead of Azure Functions and GCP with regards to programming language options.
  • However, AWS realizes that even with support for additional programming languages, there are still going to be developers that will want to write functions in languages that AWS does not yet offer. For example, what about the users who want to write functions using COBOL or PHP? AWS made a very wise move in launching BYOR (Bring Your Own Runtime) support for those users who want to write functions using languages and runtimes not provided by default. This helps to encourage further enterprise adoption — even enterprises still writing applications in COBOL!
  • In fact, Thundra also launched Java and Node.js custom runtimes along with the AWS announcement. Our custom runtime for Java decreases cold starts frequency and duration by providing a JVM with specified parameters. It also eliminates the need to change your handler. Our custom runtime for Node.js allows you to instrument your function without making any code changes by automatically wrapping your handler. Our Python Layer will come very soon with many capabilities.
  • AWS Step Functions were already very powerful for orchestrating AWS Lambda functions, especially for uses cases such as long-running tasks or tasks that run in parallel. At Re:Invent, AWS announced that you can now integrate Step Functions directly with AWS native services such as SNS, SQS, DynamoDB, and more. This means that you can now orchestrate your functions even with less effort.
  • Lastly, AWS announced other news that I believe will make AWS Lambda even easier to use and adapt. For example, ALB support for Lambda, which enables developers to set content-based routing rules for their Lambda functions and to use ALB as a direct proxy for HTTPS connections, and Web Socket support for API Gateway, which enables developers to set up two way communications with the outside world by Lambda functions. I believe both of these are and remove other barriers for those entering the serverless world.

To sum up, based upon the massive booth activity we experienced and all the serverless-related announcements, I am totally convinced that AWS continues to lead innovation and adoption in the serverless space. I am so proud to be partnered with such a leading edge organization and have already booked my calendar for the next Re:invent. I strongly recommend that everyone in the serverless community do the same!

Originally published at




Identify & troubleshoot bottlenecks in your CI workflows.

Recommended from Medium

The Data Scientist Journey: Chapter 4

Parallel Programming with Scala — Leetcode #113 “Path Sum II”

Git@Cheat Sheet..very handy..!!

Android MVVM with compression to MVP

The Joy of Building Snaps for Python Applications

What is Amazon Honeycode??

How to: EIGRP Routing Protocol Implementation & Tutorial.

On Adventure Through Shapeshifting Test Automation Strategies

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Emrah Şamdan

Emrah Şamdan

Senior Product Manager | Serverless

More from Medium

Arcadia had a little Lambda: A fleece not quite as slow

Photo of sheep grazing around an array of solar panels, located in a grassy field

Create Lambda to fetch the freeform feature flags

Building Lambda Layers in CDK

Monolith to Serverless in 10 minutes

Monolithic Service conversion to Serverless