Unstable “hanging” behavior of Cypress tests in GitLab CI

[Ponomarenko-Alice]

Hi! I’m looking for help with an unstable “hanging” behavior of Cypress tests in GitLab CI.

Problem summary

• When I run tests locally, they behave predictably: they either pass or fail.
• In GitLab CI, some e2e tests/specs occasionally (not always) hang: the status stays Running indefinitely until the GitLab job/stage timeout is reached.
• At that moment, the logs usually end with something like: Running: <spec>.cy.ts, and then nothing else appears. Because of this, it’s very hard to understand what went wrong, and the remaining specs are never executed.

Context

CI run command:

 "scripts": {
    "tests:ci": "bash -c 'source .env && cypress run --browser chrome --spec \"cypress/e2e/**.ts\" --env allure=true,host=$HOST,grepTags=\"$GREP_TAGS\"'"
    }

run cypress:
  stage: test
  extends: .default_job_template
  image: ${REPO_V2}/build/cypress/included:cypress-15.4.0-node-22.20.0-chrome-141.0.7390.54-1-ff-143.0.4-edge-141.0.3537.57-1
  cache:
    - key:
        prefix: ""
        files:
          - yarn.lock
      paths:
        - /root/.cache
        - ...
  script:
    - !reference [.add_env, script]
    - yarn install
    - yarn run tests:ci || true

Sometimes, before the hang, I can see background Chrome requests to external domains in DEBUG=cypress:* logs (for example googleapis / android.clients.google.com / safebrowsing), but I’m not sure if they are directly related.

Question

Is there any built-in mechanism or hook in Cypress that allows:

• Detecting such “hangs” (for example, based on lack of progress/events or a per-spec timeout)
• Aborting execution of only the current spec and then continuing with the next spec files within the same CI job

[MikeMcC399]

@Ponomarenko-Alice

#31018 (comment) has a suggestion about how to add spec timeouts globally.

[Ponomarenko-Alice]

Thanks for the suggestion! I’ve tried this solution, but unfortunately it didn’t resolve the issue.

In my case, the hanging is definitely caused by a TypeError occurring during test execution. When this happens, proposed hook doesn't trigger and doesn't interrupt the spec execution.
At this point, it looks like runtime TypeErrors are not being handled by this mechanism, which is exactly the scenario causing the hangs in CI.

If you have any ideas on how to handle this kind of error (or force Cypress to abort the current spec and continue with the next one), I’d really appreciate it.

Used solution:

describe("Test timeout", () => {
  context("For test duration timeout", () => {
      Cypress.on("test:before:run", function (attributes, test) {
        testTimeout(minutes(10), test);
      });
  
      beforeEach("", () => {
      ...
      });
      
      it("My test", () => {
       ...
      });
  });
});

IMPORTANT to clarify, the TypeError occurs during the beforeEach hook. I’ve noticed that errors in beforeEach sometimes behave differently.

[MikeMcC399]

@Ponomarenko-Alice

I'm sorry I can't help you more in detail. You might possibly get more feedback if you connect to the Cypress technical community on Discord and ask there.