Fix WFL deployment.

[tbl3rd]

Purpose

With the removal of the ui component, the swagger page was no longer working. This required changes to the configmap to adjust the reverse proxy to allow the swagger code in WFL to display it. In finding that solution, we also found that the swagger page could no longer allow the user to authenticate using the Authorize button. There were some missing components that the swagger page relied on to make this work.

We also found in testing swagger that there were some workloads with old values in the watchers column that would cause the api/workloads endpoint fail to return workloads.

More info provided with related issues in: #530

Related tickets:
https://broadinstitute.atlassian.net/browse/GH-1472
https://broadinstitute.atlassian.net/browse/GH-1505
https://broadinstitute.atlassian.net/browse/GH-1520

Changes

• Updated appVersion, helm chart version and wfl version
• Made changes to the helm ConfigMap to accommodate swagger being served by the api
• Liquibase changelog added to clean up watchers column in the database
• Missing files required by swagger were added to the api container

Review Instructions

• Run locally and open swagger
• Hit Authorize and you should be able to just click Authorize again and be good to go
• Once successful, hit the api/v1/workloads endpoint to make sure that results are shown and no errors occur

[okotsopoulos]

Looks good. I'd like the PR description to include ticket link(s) and a summary of the issues / corresponding fixes. Thank you!