CI crash after upgrading to angular 20 `assert.strictEqual(workerInfo.currentUsage(), 0)`

[mattlewis92]

Command

build

Is this a regression?

• Yes, this behavior used to work in the previous version

The previous version in which this bug was not present was

19.2.2

Description

After upgrading to Angular 20, our CI starts crashing when building the application for production:

❯ Building...
/home/runner/_work/clickup_frontend/clickup_frontend/node_modules/.pnpm/piscina@5.0.0/node_modules/piscina/src/index.ts:447
        assert.strictEqual(workerInfo.currentUsage(), 0);
               ^
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
1 !== 0
    at Timeout._onTimeout (/home/runner/_work/clickup_frontend/clickup_frontend/node_modules/.pnpm/piscina@5.0.0/node_modules/piscina/src/index.ts:447:16)
    at listOnTimeout (node:internal/timers:588:17)
    at processTimers (node:internal/timers:523:7) {
  generatedMessage: true,
  code: 'ERR_ASSERTION',
  actual: 1,
  expected: 0,
  operator: 'strictEqual'
}

The assertion is being thrown from the piscina dependency here:
https://github.com/piscinajs/piscina/blob/fe781d66040eb87446692a3f7056c9c2c7b84a4a/src/index.ts#L447 which was added in 4.9.0 by piscinajs/piscina#590 (angular 19 uses 4.8.0 which was why this is not a problem until now)

We patched our version of the angular CLI to bump the idleTimeout here to be 10_000, and that seems to have resolved it:

angular-cli/packages/angular/build/src/utils/worker-pool.ts

Line 18 in 2672f6e

idleTimeout: 1000,

Would you accept a PR to apply that patch? (or maybe it could be Infinity to disable this check completely? I'm unsure what the consequences of that change would be though)

I assume the issue is because our application source code is so large that it takes longer than 1s for the worker to initialise, and that timeout is hit.

Minimal Reproduction

A minimal repro is not possible as:

• The problem does not always occur (presumably because the worker sometimes initialises under the current 1s idle timeout)
• The problem only occurs on resource constrained hardware (it cannot be reproduced locally on an M3 macbook, and only within github actions)
• The problem only seems to affect large applications (100mb+ of source code). Our smaller applications do not suffer with this problem

Exception or Error

Your Environment

Angular CLI `20.0.2` + framework `20.0.3`

Anything else relevant?

No response

[sod]

Had the same. No reproduction yet though. Just wanted to voice you aren't the only one. node 22.16, angular 20.0.2

[alan-agius4]

There appears to be a race condition within Piscina where the worker idle timeout isn't being cleared before the worker is re-used and becomes active.

While we could increase the idle timeout value as a temporary workaround until it is fixed upstream, setting it to Infinity is not ideal. Doing so would prevent idle workers from ever being terminated, leading to persistent resource consumption especially in watch mode.

I opened an issue upstream: piscinajs/piscina#816

[bryanrideshark]

I'm failing to understand why they had that assert in their original logic - seems like changing that behavior would solve the whole thing.

[alan-agius4]

@bryanrideshark, replied piscinajs/piscina#816 (comment)