CASE STUDY - 2020
End to End Feature Design
Mattermost Incident Collaboration allows users to manage their incidents using a playbook and access various devops tools and interations (such as Jira and Opsgenie).
This case study demonstrates the non-exhaustive process of pioneering the Mattermost Incident Collaboration feature that would grow to provide DevOps team orchestrate time-sensitive incident collaboration workflows. The case study involves the motivation, the need, and the thought that went behind creating this tool, and the pertinent features that were included to meet end user needs.
Complete feature research and design
Maze, Figma, Html/Css, React
UX Designer & Frontend Developer (Research, Visual Design, Interaction Design, Frontend Development)
Around 10 months
Credit is due to Ian Tao (PM) and Jesse (Dev Lead) on the Workflows team, and to Andrew Brown (UX Manager)
After considerable research, and getting having a look at Netflix dispatch, along with how a lot of organisations use various Incident Management tools, it was certain that there was a huge potential market that can be provided value with the introduction of MM Incident Collaboration.
We also looked at various tools that allowed users to create workflows and manage incidents, or specific triggers to see how they were performing and what we can learn from them to help our users better.
Some of these tools included:
Things to consider for our own workflow builder
There were a lot of things we had to consider when working on Incident Collaboration. Do we want to target breadth and cater to a lot of workflows that didn't necessary fall under the category of incidents, or do we just focus on Incident Management.
Even though there was a lot of appeal initially on creating a product that can be used by anyone to create workflows for things that may not be urgent, and may aid in providing structure to their day to day activities. However, after a lot of research, talking to customers, and looking at the timeframe, it was decided that we should cater to the incident market first, and think about expanding it later if needed.
The Problem: How do we provide DevOps teams structure, automation, and intergrations along with real time communication (that they already get from Mattermost).
The Solution: Mattermost Incident Collaboration
Here are some of the key terms that one needs to understand to know how the different parts fit in the whole that is Mattermost Incident Collaboration.
A playbook is an incident template. It is created ahead of time, or even after going through a bunch of incidents and is used as a starting point for future incidents. People can create different types of playbooks depending on the type of incident.
An incident outlines the active process within a Mattermost Channel. It has a start and end time and can be either In Progress or Completed. Incidents consist of steps and are initiated with Playbooks.
An incident contains a list of steps that are to be taken to take it to its resolution. A step can be assigned a member, it can have states (Not started, In Progress, Completed), and can also have a Slash command associated to them.
A lot of time and effort went into design explorations and ideation. To present all of them would be going overboard for this case study, but to just give a hint of what I'm talking about, .
