Scale the number of Verdi worker nodes in the Kubernetes cluster

[drewm-swe]

Purpose

• This PR updates the SPS API to be able to dynamically scale the number of Verdi worker nodes in the Kubernetes cluster.

Proposed Changes

• Adds real implementation to the existing stubbed endpoints.

Issues

• Resolves [New Feature]: Refactor the HySDS workers as a Kubernetes DaemonSet unity-sps-prototype#170
• Resolves [New Feature]: Scale the number of worker nodes in the Kubernetes cluster unity-sps-prototype#171
• Gets us closer to closing:
  • As a project Manager, i want to set maximum and minimum limits on scaling for processing at venue creation time (static) unity-sps-prototype#141
    • Still need to implement scaling triggers (CPU, Memory, Job Queue)
• Resolves As a project manager, i want to set default compute types for processing nodes at deploy time, so that i can predict costs for the system (Static) unity-sps-prototype#142

Testing

• The manual verification process followed this procedure in GitBook:
  • Manual verification/testing of the SPS API on-demand prewarm-related features
• Smoke Test in MCP Test (testing with the L1B process)
• Smoke Test in MCP Dev (testing with the L1B process)

Housekeeping

• This PR is related to the following PRs:
  • Merge for 23.1 Release ades_wpst#11
  • Scale the number of Verdi worker nodes in the Kubernetes cluster unity-sps-prototype#176
• The SPS API documentation was updated on GitBook

[LucaCinquini]

Tested on existing deployment on MCP-Test