I was just catching up on the fly blog and came across the (in progress) kubernetes support. I really like the fly CLI, but I think this is a great addition to the platform to ease fears of vendor lock in.
Sometimes when fly has stability issues I consider if I should have a back up plan. I wonder if I should I have a terraform config ready that can pivot an org to AWS etc. If fly supports K8s, then I can force myself to use K8s cli/tools instead of fly cli, but stay on fly’s platform, with the confidence that in a pinch I could more easily migrate to another provider if there was a severe outage. Ironically by making it easier for me to switch I am more likely to stay.
I read the forums a lot, and I see a pretty general consensus that people here want to support fly because there are so many great things about the platform, but it can be scary when there’s an incident. And it does seem to happen on fly far more often than other providers.
So the K8s announcement is actually a great salve for that fear. Simply as a matter of best practice I was considering using the K8s toolchain against fly’s platform for the reasons stated above but then I saw the cost of a cluster is $75 USD / month.
I just want to understand that cost, its effectively a proxy API on top of normal fly usage. Why is there an additional upfront monthly fee? It feels like this is some kind of market segmentation towards larger orgs where $75 USD / month wouldn’t be noticed. But because of fly’s org model, if I have 10 orgs (and growing), in my currency that’s ~115 AUD / month * 10 orgs = ~$1150 AUD / month for a wrapper around the tools and platform I already use. It seems a bit exorbitant.
So is there some drastic hidden cost I am missing, or is this some kind of market signal not to use this for small orgs?