Improve kubernetes deployments

[BSick7]

Overview

We have full support for deploying Kubernetes apps to GKE, EKS, and AKS.
However, the process for reporting deployment information could be greatly improved.

This card entails improvements to the deployment logs and the monitoring of deployments in the UI/CLI.

Details

Here are a list of issues that could be improved:

• When running nullstone run against a k8s job, the CLI doesn't always emit the app logs. Many times, the command completes before the logs fully flush to the user.
• In many scenarios, the deployment logs don't report why a deployment failed. It would greatly improve the dev experience to enumerate all the various ways that a deployment can fail, build tests for them, and improve the deployment logs.
• It would also be helpful to report the same failure states in the Deployment section of the Monitoring tab.
• If a deployment triggers as a result of Terraform changes, it would be helpful to see those deployment logs in the same workflow.
• There should be a way to view a list of Kubernetes job executions in the UI and CLI. This should show exit status codes and similar failure info as service deployments. If there a job executions that are active, it should show up-to-date status on the lifecyle (e.g. creating container, pulling image, starting, etc.)
• We need to build a better design for the Deployment section in the Monitoring tab. It's bulky with large jobs, hard to read, and doesn't invite interaction. (Ask me for a screenshot of current state.) Each Deployment revision, pod, job execution, etc. should contain useful information to diagnose failures in a clean, concise way. Each should also provide actions and links to debugging resources. (e.g. link to commit sha in github of app code, resource in AWS/GCP console, log output in Nullstone, command to ssh, etc.)