Hey folks! I hope you're all doing well. I've dabbled a bit with Kubernetes in the past using EKS and AKS for small home projects, but now I'm working on something more serious. I plan to deploy a resource-intensive application that really strains my PC when running it through Docker. I'd love your advice on what kind of VPS specs I should be looking at. My app consists of:
- 8 Spring services
- 2 MongoDB instances
- 1 RabbitMQ instance
- 3 PostgreSQL instances
- 1 Ollama instance running Mixtral 1.5
- 1 Chroma instance
I understand it's tough to provide an exact figure, but any general estimates would be greatly appreciated! Thanks a lot!
3 Answers
Given your app is already running, you should have some baseline metrics. But if you're looking to optimize, it might also be helpful to get insights from others managing similar setups to see if they’ve found sweet spots for resources.
It sounds like a hefty workload! If you can, take a look at how much resources your app currently consumes. That’s usually a solid starting point when sizing up for a VPS. Just be aware that things might change when deploying it on a cloud service.
That was my initial thought, but it maxes out 16GB of RAM and all CPU resources, causing terrible response times. So I’m not sure that’s a viable metric.
From what I gather, Mixtral could be your main resource hog, leaving limited resources for the others. Tuning that might lead to better performance overall!
Exactly! I’m reaching out for advice from those who’ve managed projects like mine to get a check-up on my estimates.