The Fedora Community Operations (CommOps) team held a team sprint, or Fedora Activity Day, from January 29-31, 2018. CommOps provides tools, resources, and utilities for different sub-projects of Fedora to improve effective communication. The FAD was an opportunity for us to further our mission by focusing on two primary goals and two secondary goals for 2018.
The CommOps FAD aimed to carry out these primary goals:
- Pursue plan of deploying a GrimoireLabs dashboard, visualizing fedmsg data
- Launch Fedora Appreciation Week in 2018
This article explains what we accomplished in our FAD, how we have progressed since then, and what is next for the team.
Understanding the goals
The two primary goals identified for the CommOps FAD were building a GrimoireLabs dashboard with fedmsg data and launching a Fedora Appreciation Week in 2018. A secondary goal was evaluating the Fedora elections process and identifying areas of improvement.
These goals enable CommOps to create more “heat and light” (read: “contributions and exposure”) inside of the Fedora community. CommOps does this through technical and non-technical areas of contribution. Our two goals are split between technical and non-technical strategies to enable this.
GrimoireLabs dashboard: powered by fedmsg
What is it?
GrimoireLabs (or Grimoire) is a visualization dashboard tool. Grimoire offers charts, graphs, visualizations, and other ways of representing various data sets. Unlike other visualization dashboards, Grimoire is better adapted for community-related data than strictly technical data (like sensors or machine performance metrics). Grimoire features plugins for many tools already in use in Fedora, such as AskBot, Bugzilla, Hyperkitty, Mediawiki, supybot, and more.
Fedora has its own powerful data tool called fedmsg. fedmsg produces a live feed of data for various activity around the project. This includes Pagure git commits, new blog posts on the Fedora Planet, new package updates, and various other activity in the Fedora community. CommOps wants to build a data retrieval plugin for Grimoire to integrate with the existing fedmsg data feed.
Why do it?
Now, only developers use fedmsg to make meaningful conclusions from available data. Using fedmsg requires knowing Python, knowing the Datagrepper API, and creating scripts to collect and use the data. Fedora’s community is a diverse range of contributors. Not everyone is able to write fedmsg / Datagrepper scripts, but everyone is able to learn from the data available. Grimoire makes community metrics more accessible for sub-projects and teams in Fedora to better understand how they work and how they can improve.
Some of the meaningful insights we want to find are as follows:
- Determine what times (days / months) people contribute the most
- Evaluate event success if new contributors enter the community after a completed event
- Understand new contributors better and where they snap off (if they do)
- The Epic Journey of the New Fedora Contributor (i.e. using the visualizations to see pathways / successes / failures of new contributors entering the community for the first time)
How do we do it?
At our FAD, our discussion and goals focused on a plan to launch a Grimoire dashboard for Fedora. This includes a way to gather and retrieve data from fedmsg for visualization. We also identified the type of data we want to collect and how to sort it.
We identified three types of data categories of interest to Fedora contributors:
- User-based: User activity (time-wise, by activity types, etc.)
- Teams + Community Health: Different sub-projects/ SIGs in Fedora, their activity, and health
- “Fedora Land”: Broad overview of Fedora Project
More information about the types of data are found in this Pagure comment.
At the FAD, we proposed this rough timeline of order of activity for how to start the dashboard:
- February: Install Grimoire in CommOps Infra cloud
- Documenting requirements, configurations, needs for Fedora Infra proper
- March: Have a data feed loading into our instance of Grimoire (e.g. Hyperkitty / mailing lists)
- April: Continue building dashboard from sample data available
- May – July: Perceval fedmsg plugin in progress
- August: “Community preview” of the Grimoire dashboard at Flock 2018
- GSoC: Check-in on summer projects (Perceval, Happiness Packets)
- Import / normalize fedmsg data into Grimoire
- Import / normalize fedmsg data into Grimoire
- Setting up a CommOps dashboard
- November – January 2019: Setting up a CommOps dashboard
Fedora Appreciation Week 2018
What is it?
Fedora Appreciation Week is a week-long event where Fedora contributors say “thank you” to each other for the hard work put into Fedora. To support this, contributors send a thank-you email or Happiness Packet to another contributor in the Fedora community who helped or guided them. Additionally, community members can send stories about other contributors to share on the Fedora Community Blog during the Appreciation Week. A Fedora-themed Happiness Packets website could be available too.
Why do it?
Many people volunteer their available time to contribute to Fedora. Some contributors go above and beyond expectations. This event builds on top of the Friends Foundation of Fedora. Our goals are to help lower burnout rate and support positive contributions by members of our community. Organizing a week-long, virtual event to celebrate our collective achievements and accomplishments builds a positive community morale, both for our community and for each other as fellow contributors.
How do we do it?
We identified the following steps to enable a soft launch of this event in November 2018, the anniversary of the Fedora Project:
- Create template to collect contributor stories from community
- Set up Pagure tooling to accept contributor stories
- Publish Community Blog post explaining contributor stories, teasing Appreciation Week
- Mailing list / community outreach / direct pings (to build awareness for contributor stories)
- Review collected contributor stories
- Plan for preview at Flock, leading up to the real Appreciation Week
More detailed information is in this Pagure ticket.
Elections in Fedora community
Fedora elections are a key part of community engagement and participation in the Fedora community. They are a critical way to engage with the contributor community to shape direction of progress. CommOps want to find ways to make elections more interactive and engaging both for candidates and the community. Before the FAD, we had two goals in mind:
- Document and formalize our supporting role with Fedora Elections
- Improve engagement in community elections to increase participation
Fortunately, the Fedora Program Manager, Jan Kuřík, was on-site to give perspective from his position as election wrangler for the last few years.
What we learned
There were two major things we learned from talking with Jan:
- Realistic, successful metrics for a Fedora election
- Pyramid of Fedora contribution types
First, the discussion started on voter participation. Are voters disengaged? The conversation shifted to whether low voter turnout is a symptom of poor engagement. We looked at data from the last few election cycles and compared it against Matthew Miller’s State of Fedora slides. The numbers of core, active contributors in the project was approximate to the number of unique voters. Without more detailed data, we speculated whether we currently succeed at engagement.
Secondly, we learned about a “pyramid” of project contribution depth on the project:
- Bottom level: Packagers who contribute once a cycle
- Middle level: People interested in community activity
- High level: People who take part in community activity
Knowing this, we considered if messaging was an area of improvement. Why does Fedora strive to have a non-technical community? If active contributors engage with the election cycle, where could we improve and better communicate?
The discussion’s outcome played an influential role in our sub-project redefinition. Improved clarity was a highlight on how CommOps is involved with elections.
Since the FAD, the visual timelines we created are the primary measure of progress and whether we are on target to meet our goals.
Grimoire dashboard preview for Flock 2018
A live preview of a functional Grimoire dashboard is our goal for Flock 2018. With existing data retrieval plugins, we can build a working dashboard now. This lets us leverage the benefits of visualizing some data and demonstrate the value of integration from the fedmsg bus. A fedmsg data retrieval plugin is also on the list, but could be incomplete before Flock.
Since the FAD, we have a live installation in our private cloud instance. We suffered a setback by chicken pox, but remain confident on our ability to demo at Flock 2018. Sachin was prepared to work on the Perceval retrieval plugin over the summer months if we are unable to support the project with a GSoC or Outreachy project slot.
Contributor stories preview at Flock 2018
We also want to tease Fedora Appreciation Week by sharing some contributor stories at Flock 2018. How we will incorporate the stories into Flock is part of our planning for the summer months after we receive contributor stories.
So far, the template and tooling are done. A Pagure repository is set up to collect new stories, public or private. We’re preparing a call for story submissions in April 2018 on the Community Blog as our next step.
Come join us!
Does any of this work sound interesting? Curious to learn more or jump into the conversation? Leave a comment below or consider joining our mailing list and IRC channel on Freenode. If you still want to know more about CommOps, read more about us. You can also learn more about joining the CommOps team too!
Read another event report by FAD attendee Jona Azizaj.
Featured image by Justin W. Flory. Stargazing photo by Julian Paul on Unsplash.