When a page crashes with error "Page Crashed!", the browser instance is not closed properly even after the tester worker is respawned. It stucks the tester worker (it is respawned after 5 minutes of inactivity) and keeps open chrome instance hang forever.

There is no way to set some script before navigating to the configured url. This means, that there is no way to set extra headers (i.e. authentication) before navigation.

We should implement beforeNavigationScript and afterNavigationScript configuration to be able to perform said actions.

Digest: sha256:714a07652e353d2c036f19723ad30d7f7d431124e7b131d46bb539e49aa3711d
Status: Downloaded newer image for qape/qape:latest
WARNING: The requested image's platform (linux/amd64) does not match the detected host platform (linux/arm64/v8) and no specific platform was requested
exec /usr/bin/dumb-init: exec format error

adding --platform=linux/amd64 did not help.

There are several issues with puppeteer running into "Page Crashed!" error.

QApe is running into the very same issues in longer test runs. When this error occurs during Random Scenario, it causes the test run to exit as a failure, even though no problem on the tested website has been detected. QApe is also able to recover from this error by itself. For this reason, we will be not failing the test run when Page Crashed! error occurs.

This topic should be revisited, once puppeteer is more stable. Or we can consider migrating to an alternative tool like playwright.

Update the structure of the reporter event names with worker name prefix.

Examples:

runner:start -> tester:runner:start
process:exit -> master:process:exit

This should make clearer, where do the events come from.