Sumo Logic

Metrics allows Site Reliability Engineers to build powerful visualizations from unstructured data in seconds.
My Role
UX
Interaction Design
Prototyping
Team
Soraya Nunkunkit - Lead Product Designer
Bret Scofield - UX Researcher
Brian Bozello - Senior Product Manager

The Challenge

A year after launching Metrics, this product area faced low adoption. Despite training content, Site Reliability Engineers found the search language too complicated. This was a hurdle for SREs who want to add charts to dashboards and share with their team.

Early MVPs

I experimented through prototypes several approaches an SREs can take to search for content at large scale from rich visualizations to a catalog with filters. Given the speed at which SREs need to find data, a minimal approach was the winner.
When creating a chart from scratch an SRE is looking for specific metadata tags within a batch of 18 million time series or more.

New flexible search

We reduced cognitive overload by breaking apart the search into 4 components (source, metric, groups and aggregation) and presenting it in a Mad Libs style. Advanced mode is still available for power users who want to just create the search.
If the SRE cannot remember the exact name of a value, the system aids him via a reference table that list the most relevant time series. This presented an opportunity to evolve the chart builder layout so each component is visible as the SRE moves from searching to fine tuning the visualization.
Our effort has made search less intimidating which is reflected in the increased amount of searches SREs ran and charts added to dashboards.

Learnings

I learned about the importance of defining success metrics before jumping into design to anchor the exploration process and evaluating them critically too. This project went through several MVPs that offered many insights in each round but I would have like to shorten these trials to get customer feedback more quickly. It’s a craft to balance exploration and delivery.