This is a new feature request.
We would like to have the ability to set a maximum time allowed for a build.
We have a complex build (450+ projects, 2500+ tasks) and when we run our unit test and integration test suites, they can hang under certain circumstances related to external services - including but not limited to network outages. Most of the time, those interruption are only transient and re-running the build the next time will not hit those issues.
Right now, we use our CI system to timebox the time allocated for our unit/integration test process and the gradle process get killed by the CI when this happen.
Unfortunately, since gradle is interrupted that way, it gives us little information about what gradle was doing at this point. Since we use parallel execution, it's hard to understand in the log which tasks were currently being run vs which one were not. And no build scan is generated.
We would like to have the ability to give gradle an overall timeout and at the end of this timeout, gradle would interrupt currently running tasks and generate final actions (such as build scan) before shutting down.
This is a new feature request.
We would like to have the ability to set a maximum time allowed for a build.
We have a complex build (450+ projects, 2500+ tasks) and when we run our unit test and integration test suites, they can hang under certain circumstances related to external services - including but not limited to network outages. Most of the time, those interruption are only transient and re-running the build the next time will not hit those issues.
Right now, we use our CI system to timebox the time allocated for our unit/integration test process and the gradle process get killed by the CI when this happen.
Unfortunately, since gradle is interrupted that way, it gives us little information about what gradle was doing at this point. Since we use parallel execution, it's hard to understand in the log which tasks were currently being run vs which one were not. And no build scan is generated.
We would like to have the ability to give gradle an overall timeout and at the end of this timeout, gradle would interrupt currently running tasks and generate final actions (such as build scan) before shutting down.