James Gregory

I wanted to write about how using Async Iterators in Node.JS can make querying DynamoDB tables more pleasant, but I got side-tracked with explaining why pagination and is more often necessary when using DynamoDB than when using relational databases.

So instead of the thing I meant to write about, here’s a refresher on how bounded and unbounded queries are treated differently by typical relational databases and DynamoDB.

When the pandemic hit we (at my social enterprise side-hustle) quickly built a video calling web app for medical students to able to continue their “face-to-face” practice exercises, which later grew to allow actual exam scenarios for those students, and also then started supporting rehabilitation for people with traumatic brain injuries.

It’s been quite a journey and I’ve learned a lot along the way. Recently, I’ve been trying to improve the dev loop for making visual changes to a delicate part of the system.

I had one CloudTrail. Then I had many CloudTrails. Now I have only one once again, but a better configured one.

In this post I walk through my experience of migrating from a less-than-ideal hand-rolled multi-region CloudTrail setup to an single multi-region CloudTrail with a sprinkling of extra security by using a separate AWS account to store the logs.

When I was setting up Honeycomb with my Lambda functions there was something that bothered me: the OpenTelemetry config file contained my API keys and environment-specific details. I needed to keep my bundle environment-agnostic, and I really didn’t want to be committing API keys to version control. You can read more about the adventure itself in my other post, or continue reading this post for my solution.

This is a condensed guide to how I configured AWS Lambda to work with Honeycomb and OpenTelemetry.