Yup, I'm one of those people that delayed moving my developers, testing, CI/CD and Production off Docker for as long as I could.

And now, we've gotten all the value we could from Docker, Docker Swarm, Docker Compose, and its time to move on.

I'm pretty sure the journey will take most of the year, but Kubernetes has come a long way from the time we first heard of it, sitting in the Docker offices in San Francisco.

After working with Darren and the Rancher team for a year, I'm obviously going to start with k3s, add TimescaleDB, and then layer our apps over the top.

$1

 yup, talk about “it just works” – next time, need to remember the --write-kubeconfig-mode "0644"option 🙂

$1

Ah, yes, k3s does lots of things, but it doesn't give you the helm cli…

potter off to the latest release tag of helm, after finding a GH issue on k3s about the lack of helm cli support – horrible UX: download using Firefox, then…

$1

 oh, yeah, k3s – so need

$1

looks like Tobs makes a bad assumption too –

$1

yup.

tobs is running tho – so its ‘just' because I've not created a the ~/.kube/config file 🙂

$1

creating an empty ~/.kube/config isn't quite enough – but there's a PR for this issue.

build it, and use it – it works…

$1

at which point, I stopped writing, made a “Ruuvi-tag to Prometheus metrics” program, and have been watching the temperatures of our new fridge, outside and inside.

Its clearly time to redo all this using Terraform.