Monitoring

On the way to AR/VR-assisted monitoring

With augmented reality and virtual reality on the rise, much research is going in to how those technologies can support troubleshooting in complex systems. A recent CapGemini report ("AR and VR in Operations") shows some examples how AR/VR can be used to link the physical and the virtual world in automative and manufacturing industries. Why shouldn't DevOps and IT Operations benefit from advanced virtualization and interaction as well?.

When the root cause of a problem in a complex system needs to be quickly found, visualization is a key technique - for example, in showing relationships between various monitoring and logging datasources. Humans are not good at interpreting raw data, but can very quickly spot patterns in charts and discover causality between observations. Unfortunately, the monitoring systems we use today make it hard to see the big picture: screen space is limited, dashboards show only selected data, and we can't immerse ourselves into the systems landscape to find the one small thing that is wrong. So how can you easily connect with "big picture" view of an architecture with real-time information on how those components are doing?

As it turns out, it's fairly easy to build an interactive architecture model using readily available tools - a whiteboard, monitoring systems that support direct links, printed QR codes and a smartphone.

We start by sketching out the most relevant parts of system architecture on a whiteboard - that allows things to be moved around easily when the architecture changes. Each component on the board gets "augmented" by adding a printed QR code linking to monitoring data for that component - there are many free QR code generators you can use, provided that your monitoring systems support directly linking to dashboards (such as Grafana) or filtered logs.
Smartphone cameras will detect QR codes when scanning across the map and open the monitoring links they contain. Even though screen space is limited on smartphones, it's quite easy for IT admins to scan for anomalies this way.

Of course, no one will argue that a whiteboard with QR codes constitutes an AR/VR system. Over the next years, monitoring vendors will need to reconsider how they visualize data and topologies, as AIOps platforms continue to automate the repetitive work of recognizing and understanding technical issues. Meanwhile, organizations that have small independent development teams instead of a central IT Operations unit will want to consider using "low-tech" methods such as the one presented, for running their digital services with minimal downtime.