Anypoint Monitoring 2.0

→

Overview

Mulesoft is a division within Salesforce that focuses on helping large enterprise customers solve their data integration use cases. These customers typically stem from traditional verticals (i.e. banking, airlines) and possess large data systems scattered across multiple different environments / sources. Mulesoft offers a solution in the form of Mulesoft Application Networks - a Mulesoft-powered ecosystem of applications and APIs that work together to support the seamless integration of data systems (no matter their types and forms).

The Anypoint Console is the central platform built by Mulesoft to aid in the end-to-end management of customers' app networks. Users can leverage Anypoint to traverse through their entire Mulesoft journey, using a range of different tools. Anypoint Monitoring - in this respect - is a key piece in the Anypoint product portfolio that focuses on the post-deployment Monitoring & Observability aspect of this journey.

→

My contribution

Design strategy UX research Product design Usability testing

The team

1 × product manager 1 × product designer 1 × technical writer 4 × engineers

Year

2022

Process

Understanding the business impact

At the beginning of this project, Anypoint Monitoring had launched its version 1.0 a couple of years earlier. As a product, Anypoint Monitoring 1.0 did not perform very well. It was not meeting the desired business & product objectives which our senior leadership had laid out. Concurrently, it was a very engineering-driven initiative, built without a dedicated designer, and very detached from the parent Mulesoft Design System.

(Anypoint Monitoring) was not meeting the desired business & product objectives which our senior leadership had laid out.

3 example product metrics we knew we needed to improve include:

Mean time to issue detection
Mean time to issue identification
Mean time to issue resolution
Monthly user rate
Customer satisfaction (CSAT)

Understanding the primary user's context

Through a lot of discovery user interviews and surveying, our team was able to identify a major recurring problem theme with Anypoint Monitoring 1.0 - its inability to provide primary users (DevOps engineers) with network-level performance insights. Simply speaking, it did not have a performance overview dashboard for the entire network users were managing. With Anypoint Monitoring 1.0, users would have to select a single entity first (an app or API ) before they could consume any monitoring data.

At the same time, it couldn't sufficiently provide users with the monitoring data they needed for troubleshooting tasks. In the DevOps world, troubleshooting performance issues require access to 3 types of monitoring data: metrics, logs, and traces. A DevOps engineer would need to investigate all 3 categories in order to extract meaningful insights from a particular performance issue.

Anypoint Monitoring 1.0 could only offer them metrics and logs. And what's worse, they could only offer such data for single entities, not the entire network.

Design ideation & workshop

Once we had a grasp of our primary user problems and needs through UX research, the next step was to envision what Anypoint Monitoring might look like in the ideal world.

What would make our users go wow? To help figure out this question, I organized a series of design-thinking workshops. From figuring out the user's mental / conceptual model, understanding technological possibilities, assessing the market and competition, to the simple art of raw sketching and ideating. Our PMs, TC Writers, Devs, and even Sales Engineers all participated.

3 important outcomes emerged from these exercises. First, a collection of scenario storyboards that really put us in the seats of our users. Here's a sample that focuses on issue detection and troubleshooting.‍

Second, an extensive user journey map that aims to capture how different personas would work together to address various different monitoring goals. Here, the focus isn't just on the DevOps engineer, but with whom they might work in a particular organization, and what their different jobs-to-be-done are. How do these different persona goals overlap? Along these journeys, what are the UX challenges and opportunities?

And lastly, something that genuinely got us excited about: a North Star prototype. Not only was it a constantly evolving single source of truth that could help our team get aligned on the future we're marching towards, it was also a valuable artifact my PM and I used to sell the vision to leadership and customers.

Scope planning

Once we got the needed thumb-up's from the appropriate decision makers, we immediately got to the next important question: How will we get there? What will it take? What does the timeline look like? How will we phase it out? What will the milestones be? What immediate user outcomes / jobs-to-be-done do we need to solve for? etc.

We decided to prioritize the most immediately critical area that needed a solution for:

How might we help DevOps Engineers detect and troubleshoot performance issues faster?
......By providing them with a network-level performance dashboard, and a seamless yet robust filtering mechanism to drill down into the network based on important parameters.
......By offering access to traces-level data across all the entities that live within the network.
......By improving the existing Entity Details view, where users can obtain a contextual insight into the performance health of a specific entity and its impact on the overall network.

Design-test-iterate

After 4 months of planning, scoping, multiple rounds of design-test-iterate, I was able to produce a set of hi-fi design specifications / mockups that would support our near-term vision for Anypoint Monitoring 2.0. Below are some sneak-peaks.