Pre re:Invent 2023 Announcements You Don’t Want to Miss

Andy Jassy, live at AWS re:Invent 2020

AWS re:Invent 2023 is only two days away. In the lead up to the event, AWS announced several new AWS service enhancements and features that you should know about. This article contains a brief recap of 6 of these notable announcements.

#1 – Code Whisper + AWS CLI

The AWS CLI is a bit of a pain in the butt to work with. Now, with this new enhancement it will be a little bit less so!

This new feature allows you to use CLI autocomplete and type ahead code suggestions in your favourite terminal. It also offers a second feature that focuses on natural language processing and command conversion. You can use it to type natural sounding phrases like “list all of the s3 buckets in my account”, and Code Whisperer will generate the corresponding AWS CLI command for you. The combination of these two features make interacting with your AWS through the CLI just a little bit easier.

#2 – Amazon PartyRock

Amazon PartyRock is an AWS app that helps you build other apps using generative AI and LLMs. You use it by typing in a prompt into the web tool’s interface that outlines the type of app you’d like to create. For example, you can say “create a recipe generator app” that will create a series of prompts related to recipe generation. Things like asking for ingredients you have available, cuisine, time, and more.

Creative users have created a whole host of apps that you can check out including Movie Recommendations, Team Building Activities, Vacation Suggesters, and more. The tool is free to use and available today.

#3 – CloudFront KeyValue Store

Amazon announced CloudFront KeyValue store that is now available within you CloudFront functions. This means you get access to a robust and scalable in-memory that you can use in your application. The main idea is that by using KeyValue store you don’t need to set up a cache or dedicated database to store simple data. The feature is useful in many different contexts such as feature flags, A/B testing, and environment variables.

#4 – SNS Message Archiving and Replay for FIFO Topics

SNS now supports in-place message archiving and message replay for FIFO topics. If you happen to be using FIFO, you can now keep a storage log of all messages pushed to the topic and replay events from specific points in time.

This feature is interesting because the replay functionality is controlled by the subscriber, not the topic owner. This makes the feature very useful in outage scenarios where a service goes down, and you as a consumer need to re-drive your events.

#5 – Lambda JSON Log Support, Debug Level Filtering, and Log Group Custom Destinations

This announcement contains three new features. The first is that you are now able to capture JSON data from your log entries as opposed to plain text. When using JSON, CloudWatch will automatically parse your JSON objects and allow them to be queried and filtered.

The second is filter support for common log levels like DEBUG, INFO, WARN, and ERROR. Log levels like these are standard for most logging libraries. This new feature allows you to easily look through logs based on specific log levels.

Finally and most importantly, you can now specify specific Cloudwatch Log Groups for your Lambda Function. Previously, a log group would automatically be created for you for each of your Lambda Functions. This new feature means multiple different Lambda’s related to the same application can now push their logs to the same group. All logs in one group make it easier to query application issues and eliminate the need to jump through different groups to find what you’re looking for.

#6 – Step Function Restart From Failure

Step Functions are an orchestration service that allow you to build workflow oriented applications. Previously, when your Step Function failed a specific step in their execution, you would need to restart the execution from the beginning. This approach could cause some unintentional side effects. With this new announcement, Step Functions now support restarting workflows from a specific failed state.

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts