Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Date

Attendees

...

  •  Greg Hill
  •  Silas Davis
  •  Richard Berg
  •  Ryan Beck-Buysse
  •  Shawn Amundson
  •  Srinivasan K
  •  Mike Lodder
  •  Morgan Bauer
  •  Pankaj Goyal
  •  Qinghui Hou
  •  Artyom Bakhtin
  •  Matt Sykes
  •  Brett Logan
  •  Dipti
  •  Ry Jones
  •  Dave Huseby

...

Agenda

Completed Work

  •  David Huseby  move the initial CI/CD document from the TSC space to CI/CD space.
  •  Gregory Hill  add details, including costs, of the Burrow CI/CD pipeline in the existing setups page.
  •  user-b2fb8  get the actual cost of the current Jenkins and Gerrit CI/CD setup used by Fabric.
  •  David Huseby  initial requirements gathering document set up.

Outstanding Work

  •  Shawn Amundson  add details, including costs, of the Sawtooth and Grid CI/CD pipelines in the existing setups page.
  •  David Huseby  look into whether Kubernetes can utilize crowd-sourced computers to extend the cluster.

Discussion

  • Talk about how we want to do requirements gathering.
    • Should we break up along teams and report back?
    • Should we assign one or two people to study each team's solution and report back?

...

  • Dave space set up and requirements docs are up.
  • Ry: we're budgeted to spend about ~$120k/year and we're currently spending about $60k this year so we do have some money to spend for some exploration with a parallel system.
  • Dave: talking about requirements.
    • Kubernetes concerns:
      • Sawtooth
        • assumes a docker runtime under the docker building process.
        • also docker compose.
      • are any projects using Docker swarm or any other competing technologies that they would have to move away from?
    • Other platforms to look at.
      • CNCF is using Prow. What is Prow?

Action items

  •  Ryan Beck-Buysse look into Sawtooth's existing pipeline and identify any pain points for moving existing infra to a kube cluster.
  •  Srinivasan K  figure out if any of our teams are using Docker swarm or any competing products.
  •  David Huseby  investigate kube cpu credits, memory limits, cpu limits, and load balancing across namespaces.