Comet: Sprint 13 (March 2, 2022 – March 15, 2022)

Note: This session was not recorded

What was the goal of Sprint 13?

  • Documentation of work done already in the work cycle. Leave things in as clean a state as possible before we break for several months
  • Set up meetings to address key questions and concerns before we break
  • Complete Tidewater OAI implementation by adding complete support for Oai ‘sets’

What is the milestone that this sprint is supporting?

Focus on documentation for all apps and setting up the initial steps to “publishing” in Comet to other services

Accomplishments of Sprint 13:


  • Machine-Readable Metadata meeting
  • User stories for publishing meeting
  • Discussion of needs for deploying multiple applications for review deployment, similar need for local development
    • Outcomes:
      • Rewrote CI/CD configuration for Superskunk, Comet, and Tidewater to use Gitlab Rules syntax and allow deploying any of those applications in a Merge Request by adding the name of the application as a Label to the merge request
      • Use of mmock for a mock service api that can stand in for Superskunk while developing tidewater locally. This is working well, unfortunately we have an issue with using it in the Gitlab CI/CD environment which will need more consideration and possibly a workaround


  • Finished documentation for deploying RabbitMQ to campus clusters


  • Documented access control plan using HTTP Message Signatures, which is currently an IETF HTTP Working Group draft
  • Support added in Helm chart to associate an RSA public key per consumer application so that Superskunk can verify the identity of the consumer requesting content


  • Oai Sets are now integrated
  • Helm chart now fully documented, including the Tidewater consumer script and deployment
  • Support added in Helm chart for having an RSA keypair used for signing requests to Superskunk for content. This will be used by the consumer script as it responds to events from RabbitMQ


  • Documentation written for creating listeners which publish to RabbitMQ. Core methods used in this process now have documented descriptions and examples, when relevant
  • Added the ability to unpublish an object that has already been published. Currently we publish only at the collection level, but this unpublish ability now means that there is the granular ability to unpublish individual objects within a published collection.

Surfliner Schema:

  • Setup CI pipeline for surfliner_schema
  • Large refactor to support multiple readers in preparation for using an m3 metadata model for Comet

Did we do everything we set out to accomplish?

Not quite, but we're almost there!

What’s next?

  • April 13 - 26, 2022: Off-cycle sprint for Starlight work
  • June 8 - August 2, 2022: Spring/Summer Surfliner work cycle

GitLab link: