Fuck the centralized cloud why does my job require it
Also “everything is so reliable in the cloud everything is better in the cloud” my ass, not even the Microshit status page shows any error, but just look at downdetector and it becomes obvious
Fuck the centralized cloud why does my job require it
Also “everything is so reliable in the cloud everything is better in the cloud” my ass, not even the Microshit status page shows any error, but just look at downdetector and it becomes obvious
I’ll confess that I’m only passingly familiar with kubernetes, does it allow for this?
Kubernetes doesn’t care how many servers and from what cloud provider you use, if you add 20 AWS, 20 Azure and 20 google cloud servers and set up a master (or more) on each provider with the rest being workers, then your servers are completely replaceable as long as the rest can take the load.
The only complication is the client connections. You’d have to split your external ips across all three providers and across countries/Geo regions.
Worse case then is that you lose AWS/Google/Azure and with it the regions that used the AWS/Google/Azure ips, then you’d have to failover DNS to the remaining IPs.
In short: It’s possible and the main problem is the user being able to connect after fail over after the original IPS change.