Improve process handling and show stdout/stderr in portforward-monitoring-satellite-harvester.sh

[mads-hartmann]

Description

This improves process handling and the debugability of the dev/preview/portforward-monitoring-satellite-harvester.sh script by relying on xargs for process management.

It also simplifies the output slightly.

Related Issue(s)

Fixes https://github.com/gitpod-io/ops/issues/1947

How to test

I ran the following script in a workspace on a branch that had a VM /dev/preview/portforward-monitoring-satellite-harvester.sh and tested how it dealt with

1. The Prometheus pod not running. Previously the script would just silently fail, now it shows an error message you can use to debug:

   E0504 09:07:39.940182   52166 portforward.go:406] an error occurred forwarding 9090 -> 32001: error forwarding port 32001 to pod 49f246b3afe6a6c7eccf0b2c62d8f2922e534edd41c38f2813654f8452e658ce, uid : failed to execute portforward in network namespace "/var/run/netns/cni-4048ab39-6cee-d3e3-880c-834fa272a0f8": failed to dial 32001: dial tcp4 127.0.0.1:32001: connect: connection refused
   E0504 09:07:39.940541   52166 portforward.go:234] lost connection to pod
   

2. Running the script in many terminals to ensure it deals with existing processes or zombie processes.

Release Notes

NONE

Documentation

N/A

[liam-j-bennett]

🛹

[mads-hartmann]

@liam-j-bennett Thanks for the review! I had to rebase on main but the diff is the same and I just verified the script so I'll merge once CI passes ☺️