Very cool, so you probably do a lot with Kubernetes and the like. Ever used AWS Redshift? I'm looking into it for a future project.
I'd describe my current role as 45% DevOps/AWS Architect/Docker drone, 45% firmware/SITL/HITL, 5% front-end and 5% physics/game engine modeling. I really like it because I get a combination of DevOps, web services, low-level and high-level development.
Front-end development (especially js) does not come as naturally to me as back-end or hardware/low-level work. It's probably my biggest weakness at the moment.
That's good though. We need more like that - less python surfers. I've transitioned from QA so I'm still in junior mode, but getting there. Sounds like you got your hands full
Never used Redshift. I'm not AWS certified, but I've found some of that space to be a bit convoluted. Last week we had a hackathon, and I was trying to work on a serverless deployment of our application. I wanted manual remediation when some conditions were met (like if a CC was rejected, or some endpoint was exhausted [like space limits or time limits or money limits... whatever]) and was trying to use stepFucntions with some Lambda calls to deploy an email to the end user should we exceed or bork one of those limits/fields. It was a major pain in my dick. Took me almost two days to setup that function alone. The proprietary "language" they were using made it way more confusing than seemed necessary.
Like I said, i'm still a baby in this space, so it's probably just me. But it was tough for sure (maybe just pebkac).
Kubernetes (and docker to a smaller extent) is great - I can deploy everything in little lightweight containers/pods and scale infinitely easily for next to nothing (for kube, anyway). A lot of my testing revolves around node farms that would otherwise just be non-feesable to run locally. I can keep current with develop builds with a simple kube script - that's the play. No manual installs, no real upgrading - just ssh + deploy via the pre-configured script.
I've spun up demo/staging environments for the whole company to hit whenever they want to test or try something. Since there's generally between 10-15 people in there a day, we spun up some grafana/prometheus instances there as well. We're pulling tons of valuable data out from it. It's not yet *perfect*, but it's getting there.