Improve Logging in Data Science Pipelines pods

[rimolive]

Every time we need to enable debug logging in DSP components (API Server, persistence agent, scheduled workflow), it is a pain to make it work. We need to enable logging through manifest parameters and inject through env var in all pods to enable debugging.

[DharmitD]

The initial scope of the task was around just making manifest changes to our midstream repo. On further investigation, we figured that we would need to make changes to the code base as well (main.go for ever component), irrespective on whether we choose to use logrus or glog/klog packages for setting logging levels.

For example, Persistence Agent doesn't have the glog package imported: https://github.com/opendatahub-io/data-science-pipelines/blob/master/backend/src/agent/persistence/main.go#L26
We would need to import the package, and also add some code the way we've added in this PR, to make sure we have a log-level cli parameter, through which we could pass a value and set the variable for the desired log level.

Similarly, we would need to update the code base for APIServer and Scheduled Workflow components as well. Hence, this change needs to go into the upstream kfp-tekton repo first.

[DharmitD]

Closing this task since this change needs to go into upstream first. Created #158 to address this for v2.