we have Composer 2.6.6(Airflow 2.5.3), and a job VANI-UEBA3 which is running on Dataproc Serverless Batches … the job runs through fine (as shown on the Dataproc Serverless UI),
but the composer UI is showing errors very often – error: Task exited with Negsignal.SIGKIL – which indicates the process is killed due to more resources being used by the task
snippet of the error :
[2024-06-10, 00:41:54 UTC] {credentials_provider.py:353} INFO – Getting connection using google.auth.default()
since no explicit credentials are provided.
[2024-06-10, 00:44:32 UTC] {local_task_job.py:212} INFO – Task exited with return code Negsignal.SIGKILL
[2024-06-10, 00:44:33 UTC] {taskinstance.py:2599} INFO – 0 downstream tasks scheduled from follow-on schedule check
However, when i go to the Dataproc serverless UI, the job has actually succeeded.
Also, the job is running on the Dataproc serverless cluster .. not on the airflow cluster so ideally there should not be an Negsignal.SIGKIL error after the job is completed. (if the job is not launched, that might be acceptable error IMO)
Any ideas on what might be causing this issue and how to fix this ?
tia!