You lose some of the flexibility with docker when you are persisting data (more complex to use Docker Swarm) but it’s pretty easy on a single node.
You can create volumes, which can be backed up, but I prefer to just map a path for each container to the host’s file system. It’s a one line add. Then I just have to back up one parent folder and all my persisted data is backed up.
Yeah I've tried with both docker & k8s and both are a pain. As well, my developers don't load test or provide information to help set mem/cpu probe constraints so it's all guess and check 🤨
I’m facinated by k8s, but don’t know enough about it, and haven’t really had a reason to play with it, since vertical scaling has been much simpler for the stuff I’m running.
Sounds like you’re working with some pretty high throughput services. Hopefully it’s still an interesting problem. I feel like things get much harder when they quit being interesting and become a chore.
Lol yes and boss man just wants to know the next steps... like it's some linear issue. And unfortunately it's not interesting Lolol. Thankfully though, it's a complete side project!
