I’ve had a series of cronjobs run on google container engine and I’ve been hit with an occasional SIGTERM to one of my pods triggered by my cronJob for unknown reason:
I’ve only been able to find this out as Sentry logged my tasks receiving a SIGTERM. Unable to find the pod itself that was terminated even with kubectl get po -a
I’ve looked at the kubernetes nodes (which have autoscaling ability) and none of them have been drained and deleted. The ages of all the nodes is still the same.
I’ve also set the concurrency Policy: forbid (does kubernetes check for concurrency after it has initiated a pod or before it?)