Building the WebUI uses too many resources in a limited resource environment

[d-ylee]

Description

In our k8s cluster, our resources are limited. When deploying the WebUI with the official rucio/helm-charts, the pod that builds the webui crashes. The error in the logs is:

uncaughtException Error: spawn /usr/bin/node EAGAIN
  at ChildProcess._handle.onexit (node:internal/child_process:286:19)
  at onErrorNT (node:internal/child_process:484:16)
  at process.processTicksAndRejections (node:internal/process/task_queues:82:21) {
    errno: -11,
    code: 'EAGAIN',
    syscall: 'spawn /usr/bin/node',
    path: '/usr/bin/node',
    spawnargs: [
      '/opt/rucio/webui/node_modules/next/dist/compiled/jest-worker/processChild.js'
    ]
  }

This could be due to the pod requesting more processes or threads that is available.

Motivation

Some environments where the WebUI will be deployed in have limited resources, which would need to be accounted for, especially since the WebUI needs to be built every time it is deployed.

Change

According to https://kdaws.com/learn/fixing-next-js-error-spawn-errno-11-on-cpanel-and-plesk-with-cloudlinux/, there is a configuration in next.config.js that can be set:

module.exports = {
  distDir: 'build',
  experimental: {
    // This is experimental but can
    // be enabled to allow parallel threads
    // with nextjs automatic static generation
    workerThreads: false,
    cpus: 4
  }
};

In my use case, setting cpus: 1 and manually running docker-entrypoint.sh in our pod allows the build to work.

Additional Information

• It is related to the Storybook?
• It is related to the GitHub Actions?
• It is related to the developer tools?