I want to know some data about Google Cloud instance and load balancing

[aki1]

@Lauren_vdv

@SebastianG 

@sebastianga 

Google Cloud Compute Engine: the maximum number of connections under different instance types, the number of new connections per second, (QPS queries per second, check again if there are resources)

Google Cloud Load Balancing (Google only divides types, not specifications): maximum connections, new connections per second, (QPS queries per second, check again if there are resources)

If you can give me the source of the results,I will very graceful to you

[SebastianG]

Hi @aki1,

I think we need a bit more details about what you're actually trying to achieve with this.

Compute Engine / VMs have no notion of queries, it's just a Virtual Server connected to a network, in this case Google's network. You can use this internally or expose it externally / internet. Either way this is just a VM.

If you intend to install something like Apache, NGiNX, Node etc. on this VM then that's a different thing. I mean you'll have to do the math of how many connections you're expecting, each of these will consume a bit of CPU and a bit of RAM, so this have to be known beforehand if you're going with this approach.

Second approach will be to make use of the Autoscaling service based on metrics which can be CPU, RAM and even connections with a bit of work. This will sit behind your "fleet" and will scale it up and down accordingly. Here you'll find some details about this: https://cloud.google.com/compute/docs/autoscaler

No idea about the LB limits to be fair, there must be one in place and I think you can raise a support ticket with GCP and ask about this. Honestly I doubt you'll hit that limit, but if you do then I'm sure GCP will help with that.

There are a few options, being a managed service like GCP or you provision your own custom LBs. Being already a GCP customer I'd say talk to them, most likely they'll help your with a professional managed solution.