Too many memory used while startup nebula-studio with container on server with lot's CPU core

[matrixji]

Describe the bug (must be provided)

I using Kubernetes to start nebula-studio, and set limits with 2G memory, it always got OOM Killed on startup.
I then check with docker-compose way, it starts many node worker processes (depends how many CPU core's we have), as I using a server with 72cores, and each node work process may need about 50MB memory, so totally, it requires about 3.6GB memory, then I increase the memory limits in Kubernetes deployment, it's OK.

Your Environments (must be provided)

• OS: Linux
• Node-version: N/A using official container
• Studio-version: official container image: vesoft/nebula-graph-studio:v3

How To Reproduce(must be provided)

Steps to reproduce the behavior:

It should be clear in the above description.

Expected behavior

I think at least for a startup there is no need for so many worker processes.
Maybe a suitable startup configuration --workers=3 like a non-container startup is a simple solution.

Contents in file: package.json

{
"start": "egg-scripts start --daemon --title=egg-server-nebula-graph-studio --workers=3",
"docker-start": "egg-scripts start --title=egg-server-nebula-graph-studio"
}

And if possible, we'd better calculate the worker's number, by checking below things under container:

• How many CPUs does the node has, from nproc
• How many logical CPUs avaliable from cgroups, checking:
  • /sys/fs/cgroup/cpu/cpu.cfs_period_us
  • /sys/fs/cgroup/cpu/cpu.cfs_quota_us
• How many memory available from cgroups, checking: /sys/fs/cgroup/memory/memory.limit_in_bytes

Additional context

N/A

[hetao92]

merged