On kbuild exception, store and send to API recorded stacktrace for easier troubleshooting and error reporting

Right now stacktrace is lost as soon as pod will be cleaned by k8s.
As half-solution kci-k8swatch.py from deploy will record logs, but it is not ideal.
We have already trap for exception, so maybe we can just add some basic function to send it somewhere, or schedule to send over another pod.

Start by reading kci-k8swatch.py in deploy and tracing the existing exception trap. Determine where the stacktrace can be retained after pod cleanup and how it should be sent to the API or another pod; done means the recorded stacktrace remains available for troubleshooting and error reporting.

Written by the indexing model from the issue text.