AWS Deadline Cloud Logs

Efficient render error troubleshooting experience for faster and better 2D and 3D digital content deliveries

Troubleshooting errors takes up time to work on and deliver content creation!

In a fast-paced creative production world, digital content creators are consistently under pressure with tight deadlines and high expectations. In order to generate the final images or videos of digital contents, they need to go through a process called rendering. Not only it's time extensive, but also its debugging process can take up even more of the creator's precious time.

Teams at AWS Deadline Cloud developed a way to optimize render time for its cloud-rendering service. By working with stakeholders and internal customers, I shaped and launched the core log experience that would help creators pinpoint errors more quickly and easily so that they could focus on their assigned creative tasks.

Timeline UI that summarizes a wall of plain log texts

My interviews with customers revealed that it takes efforts to locate and put together helpful information about what caused errors, especially when not all warning messages are actually related to the cause.

An engineer and I came up with a timeline UI, which summarizes a chronological order of render events. It allows customers to grasp the contexts and jump directly to a section that's actually related to the cause without having to scroll through walls of texts.

Scrolling through a timeline UI

Streamlined log comparison

Customers often compare a render job with similar or previous versions in order to investigate any differences that may have caused an error.

They can now pull up a summary table of previous renders that allows to compare between versions at a glance before they decide to investigate further.

Interaction of opening a summary table of previous task runs

They can also toggle on “View logs for all tasks” if they want to compare their render to similar ones that happened during the same render session.

Interaction of showing and hiding other tasks happened in a render

Progressive data display to save cognitive loads and software performance

Through multiple rounds of testing and user feedback, I prioritized the essential information that would help customers make most of the default view. By reducing the amount of displayed information, the app minimizes numbers of data fetches and improves its performance.

Additional information can be a computing worker’s specs or a link to its logs.

Impact

🏆

Achieved 2 awards (Show Product of the Year Award in the Cloud Computing category and a TVBEurope Best in Show Award) from National Association of Broadcasters (NAB)

📰

Featured in 23 unique English-language articles (including an exclusive interview with Variety) within 1 month of launch

💬

Received positive reaction from stakeholders and customers:

...we gave the demo to [a customer name] and it was just awesome again. We ended up debugging an unexpected issue (rendering, not in the service) live and the UX of pin pointing what was wrong was sooooo slick…

Along with the customer’s excitement, stakeholders highly regarded my prototypes. Visualizing interactions helped align everyone’s understanding of complex engineering concepts, synergize discussion, and test our hypothesis with customers. They also complimented the 2x2 effort-and-impact matrix workshop. Prioritizing helped us focus on what's most important when everything felt important.

2x2 impact and effort matrix exercise board that my team used to determine features to include in the service launch.

Team

Partners and collaborators

Awards

Press

Next case study

Edison™ Design System

Launching a multi-award-winning design system for creating consistent, modern, and intuitive health software