I’m moving a pretty heavy inventory over to K8S - it’s a mixed-bag portfolio for a SMB marketing agency… About 300 legacy WordPress, Laravel, Node, and Golang apps, all with some sort of DB backend (or several).
I started with a Rancher RKE cluster backed by droplets and do-csi driving NFS for RWX. It was gold in preview with several apps running, but at about 15/300 apps and 8 nodes, I started running into sporadic fatal errors. Mostly nginx upstream timeouts between service and pod or pod and other-pod (db). Sometimes it’s still gold, sometimes it’s garbage. Sometimes it’s sticky on a browser session across apps, sometimes it’s everyone hitting a specific app - IMPOSSIBLE to trace/debug 🥵
I suspect it’s my fault - I likely slightly mis-configured the cluster from the start :/
So my question: I’m considering trading the RKE cluster for DOKS and (potentially) managed DBs (I trust DO enough ☺️). Anyone running similar inventory on DOKS? Can you speak to the overall cluster-level stability|reliability|resiliency? Of course, app-level trouble is on me. But the “managed” in managed k8s - what’s the word?
These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.