Building point-to-point integrations to glue systems together has always taken a big effort to build and maintain in large organisations. EventBridge Pipes make this a lot easier, with added abilities to filer with pattern matching or enrich with Lambda, Step Functions or APIs. We've got some cron based tasks and manual ETL workloads we'll be porting to pipes in the next weeks.

X-Ray is a crucial tool in our arsenal for debugging and optimising distributed Cloud-Native applications. SQS is also a key ingredient to highly scaleable systems, yet support for tracing across SQS to Lambda invocations was limited in the past, creating a blind spot for teams combining the two. We're very excited to see SQS and Lambda X-Ray trace linking giving improved observability with AWS native tools.

EventBridge forms the backbone of many of our Event-Driven Architectures, and the rules that define triggers are key to building event based interactions. Rules can be challenging to write, but can easily be built in the console. EventBridge now supports generating CloudFormation from rules built in the console, as well as automatically generating rules from the schema.

It's great to see the developer experiance for teams building with EventBridge continue to improve, enabling a truley Event-Driven future.
Keep it up EventBridge team!

It always good to hear more Serverless services getting announce but it's becoming harder every year to define what Serverless means. The new Serverless OpenSearch offering charges a minimum of 4 OCUs ~£600 a month. Scale to zero and pay-per-use were once fundamental to what it mean to be Serverless, but it's great to have a "more Serverless" version of OpenSearch.

The benefits of IaC don't need explaining, especially as we default to a much larger and dynamic number of development, testing and production environments. The introduction of IaC for Dashboards means we can define templates that can be deployed per environment, with version controlled history.

Lambda SnapStart is a huge win for latency sensitive functions built in Java. By taking a "snapshot" of the initialised execution environment, mitigating start up latency inherent in JVM based execution. This is a big when, especially for Java applications using heavy frameworks like Springboot.

Leveraging Firecracker Snapshots under the hood this can be enabled directly in the console, at no additional cost.

We're interested to see how such snapshotting can aid functions in other runtime or requiring a large ML model as data.

As regulations have been created to protect personal data companies have tightened up a lot of data storage and access, but logging has often been overlooked. Luckily CloudWatch now supports data protection policies that scan log data in transit for sensitive data and mask sensitive data that is detected. Teams often have internal policies for what they do and don't log but developers do make mistakes. Moving this to defined data protection policies removes another aspect for developers to worry about. It will also allow users with the right IAM policies to view the unmasked data which can be useful compared to not logging PII at all.

Usually when I'm looking to co-ordinate resources between accounts I look for a choreography tool such as EventBridge instead of an orchestration tool like step functions. It good that it's easier to access cross account resources but we'll see if this changes the ways we build systems.

I see too many teams defaulting to standard workflows and not leveraging express workflows where it makes sense. Often it's a simple as they prefer the observability features from standard workflows so this is a big step in the right direction!

Lambda extensions have been enabling third party observability tools to tap into function logs for a while now, improving application observability for Serverless applications. By expanding access to traces and metrics thrid party tools can now more efficiently provide rich performance context with reduced latency.

Developers do love Dark Mode!

The aleios colour scheme is a little dark mode inspired too 😉

Some data just makes more sense on a graph database but I still see most Serverless first companies defaulting to DynamoDB. Hopefully as we see more Serverless Database options we can go back to picking the right tool for the job!

For those looking for a Serverless SQL database seeing CloudFormation support on Amazon Aurora Serverless v2 is a step in the right direction. There are still some missing pieces, such as scaler to zero, before I go back to SQL databases over DynamoDB.

We've all had that moment of panic after releasing a database engine update. It's a lot less stressful to promote to a staging environment.

AWS AppSync now supports JavaScript resolvers for Pipeline Resolvers and Functions, removing the need to learn VTL. In the past I've seen many teams revert to using Lambda functions for resolvers instead of direct integrations to avoid learning VTL. With this announcement developers can now have their familiar JS syntax, linting, and testing tools. Being able to use eslint and jest with our resolvers is a big improvement for DevEx. I can see a lot more AppSync API's going lambdaless with this introduction!

It will interesting to see if teams set up a transpilation workflow to write their resolvers in TypeScript or stick with vanilla JS.

Suffix filtering support allows us to remove manual filtering in services that are triggered, reducing code to manage, and invocation costs. One change we're rolling out across projects involving file uploads is file extension filtering, triggers for PDF only uploads will be much easier now!

The ability to do OR matching and eliminate case sensitivity further reduce custom code filtering.

Great to see our favourite Serverless service continue to improve and build the foundation of Event-Driven Architecture.

EventBridge has become the backbone of many Serverless Event-Driven Architectures. The addition of EventBridge Scheduler enables cron, fixed rate, time window and one-time triggers. Outside of classic cron jobs this is a huge win for application developers. For features like reminders and notifications it's going to so much simpler to create one-time scheduled events, say goodbye to cron based database scans.

It's great we now have access to the latest LTS version of Node on Lambda, bringing the latest version of the underlying V8 engine and global fetch by default. It's also a great time to make the upgrade to v3 of the AWS SDK which will now be packaged with the new Node 18 runtime.

And, as always, the reduced management of automatic updates and patching is one of the many reasons to sleep better at night thanks to Serverless.