src: optimize usage of memory api

[theanarkh]

1. constrained_memory maybe greater than uv_get_total_memory so we should take the minimum value.
   And uv_get_total_memory maybe 0.

2. We should always report constrained memory in the report even though it is zero.

cc @Asaf-Federman

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• documentation is changed or added
• commit message follows commit guidelines

[nodejs-github-bot]

Review requested:

• @nodejs/config

[Asaf-Federman]

AFAIK, constrained memory can be greater than the total physical memory due to swaps/overcommitting (which usually are set explicitly).
Any reason for taking the minimum between the virtual memory and the physical memory?

[codecov]

Codecov Report

❌ Patch coverage is 20.00000% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.04%. Comparing base (04c5a18) to head (e653171).
⚠️ Report is 58 commits behind head on main.

Files with missing lines	Patch %	Lines
src/node_options.cc	0.00%	3 Missing and 1 partial ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #59260   +/-   ##
=======================================
  Coverage   90.04%   90.04%           
=======================================
  Files         648      648           
  Lines      191200   191201    +1     
  Branches    37472    37473    +1     
=======================================
+ Hits       172160   172170   +10     
+ Misses      11665    11625   -40     
- Partials     7375     7406   +31     

Files with missing lines	Coverage Δ
src/node_report.cc	93.36% <100.00%> (+0.14%)	⬆️
src/node_options.cc	84.40% <0.00%> (-0.29%)	⬇️

... and 29 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Asaf-Federman]

related to #59227

This does not seem related

[himself65]

yeah, I misunderstood

[theanarkh]

AFAIK, constrained memory can be greater than the total physical memory due to swaps/overcommitting (which usually are set explicitly). Any reason for taking the minimum between the virtual memory and the physical memory?

It is consistent with this. But i think you are right that the virtual memory can greater than physical memory. @bnoordhuis Hi ! could you please help take a look at this ? Thanks !

[Asaf-Federman]

As you correctly noted, uv_get_total_memory returns 0 when its value is unknown. This might become problematic if uv_get_constrained_memory is defined but uv_get_total_memory is not, as the min function would incorrectly return 0 despite a valid constraint being available.

[theanarkh]

Yes. It seem we should remove std::min in SetIsolateCreateParamsForNode.

[Asaf-Federman]

AFAIK, constrained memory can be greater than the total physical memory due to swaps/overcommitting (which usually are set explicitly). Any reason for taking the minimum between the virtual memory and the physical memory?

In summary, uv_get_constrained_memory reports the memory limit directly from cgroups—the Linux feature that orchestrators like Kubernetes and runtimes like Docker use to isolate resources.

The function correctly reflects the container's configured limit, not the host's physical RAM. This behavior is essential for advanced techniques like memory overcommitment, which carries a significant risk of performance degradation if mismanaged.

Given these trade-offs, I'm +0 on changing this feature.