Lark Sprint 4 Review Recording: April 5, 2019

What was the goal of Sprint 4?
The sprint had two main goals: First and foremost, we set out to complete cloud-native deployments Lark into both campus’s environments. Second, we wanted to take our first steps toward configurable metadata models. This a key design goal and deliverable for the product.

What is the milestone that this sprint is supporting?
The Milestone 1 is to have an initial stand-alone platform where we can create, edit, and retrieve authority records.

Accomplishments of Sprint 4:
Deployment - We achieved repeatable deployments Lark using Helm and Kubernetes. The deployment configuration is usable by both campuses.

Stable development and staging deployments are up on UCSD infrastructure at and  UCSB deployments have been tested on temporary AWS servers, but stable services are depend on ongoing work to deploy Kubernetes and Rancher locally.

Configurable Models - Initial configurable modeling is finished for Concepts, and in-progress for Agents.

Documentation - We took a pass through architecture documentation.
We also added some design work on the documentation site at

Did we do everything we set out to accomplish?
UCSB’s deployment is still pending. Though we anticipate a single Lark production deployment, shared deployments are a project-wide goal, so we consider this a gap.

The team operated at reduced capacity this sprint, owing to conference schedules (LDCX) and community work (Hyrax, Wings).

GitLab link: