fix: ensure that the correct number of CPU cores is obtained within t…

[isbasex]

PR Description

Fixes Incorrect CPU Core Count within Docker Container

This PR contains a:

• bugfix

Motivation / Use-Case:

This PR addresses a bug occurring within Docker containers due to an incorrect assumption about the available CPU cores. In the original code, os.cpus().length was used to obtain the number of cores. This works fine in most scenarios, but within a Docker container, this would return the number of cores available to the host system, not the limited number allocated to the container itself. This discrepancy leads to creating an excessive number of concurrent threads, which can ultimately result in an out-of-memory (OOM) error.

This PR corrects this issue by ensuring that the number of cores reported matches the actual number allocated to the Docker container, preventing the creation of too many threads and subsequent OOM errors.

Breaking Changes:

This PR does not introduce any breaking changes.

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: isbasex (53039b2)

[alexander-akait]

I don't think it is a good idea to read system files https://github.com/xiaoxiaojx/get_cpus_length/blob/master/index.js#L21

[alexander-akait]

Just wonder, is a problem happens for all docker containers and for all versions?

[isbasex]

In our testing, all Linux containers have this issue.

Currently, no better method has been found. If this method is not used, would you consider not setting the concurrency quantity completely according to the CPU cores by default?

For example: Math.min(4, os.cpus().length)

I think this fix is very important for robustness, especially for large projects.

[alexander-akait]

I think this fix is very important for robustness, especially for large projects.

I agree with you, just want to find the better solution. How we can detect is it a docker container? Maybe we can disable/improve logic without reading system files.