James Gregory

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.

I have several existing Lambda functions which are all built on Node.js which I wanted to connect to Honeycomb. I spent some time over the weekend working through it, and this is my stream-of-consciousness.