Analytics company New Relic is currently in the process of rolling out New Relic Explorer to its Full-Stack Observability customers. With the new visualisation capabilities hitting general availability this week, DevClass sat down with the company’s EMEA CTO, Gregory Ouillon, to get a closer look at the changes.
In 2020, New Relic said it was restructuring its observability platform into a Telemetry Data Platform for collecting operational data, and was adding Applied Intelligence for machine-learning driven capabilities as well as a Full-Stack Observability product. Explorer advances the observability tooling by providing users with an instant visual representation of their setup’s health.
“New Relic Explorer relies on the concept of entities,” Ouillon says, pointing out that in New Relic “anything that reports telemetry is an entity” from host, to application, middleware component, or a cloud platform as a service component.
Getting to grips with how these interact can be quite a challenge, especially with more complex microservice-based systems, which is why Explorer introduces a Lookout and a Navigator view.
According to Ouillon, Lookout uses the “golden signals” of monitoring, such as traffic, latency, and errors, to calculate which elements of your tech estate have the highest throughput and mark the ones whose response time has changed the most in the last couple of minutes, since these are often the ones to watch.
Results are shown in a circle diagram with larger diameters signaling higher throughput while colours are meant to draw attention to elements with decreased or increased resource usage. Having this is important to Ouillon since “visualisation is much faster in terms of interpreting versus text.”
Users of Grafana and Co might be familiar with the notion, though the combination with under the hood statistical analytics to reflect change is meant to give New Relic users an edge over other more traditional visualisation tools. From the circle diagram, teams can then start to dig deeper and get performance related metrics, look into the system’s history for anomaly detection, and check if the currently observed behaviour might have anything to do with a different entity.
“The system,” Ouillon says, “will tell you out of the box ‘Oh, by the way, you had this anomaly with your web portal, but we see the same pattern with another service.’ So we are able to surface automatically, in real time, correlations between various systems and services, which allows me to understand correlations and maybe go and deep dive into another part of your architecture, because it might be responsible for the issue.”
New Relic Navigator goes a similar route, but instead of focusing on changes (as Lookout does), it provides an overview of all the entities in a company’s system. Since entities also comprise units like containers and services – of which there can be many across accounts – Navigator displays single units as hexagons (to fit in more), using a traffic-light colouring system to find the most pressing issues quicker.
“In a few clicks, [users] can understand the blast radius of their issues, they can correlate those issues, and then they obviously can get detail, if they click on any of these honeycomb hexagons, and can see information about the alerting with the tags associated with that entity and key metrics of performance.”
Besides entity impact, Navigator also looks into entity relations, surfacing information such as which hosts support a specific service or which workloads depend on it.
New Relic Explorer has already been rolled out amongst Full-Stack Observability customers, but is meant to be generally available for new customers by 24 February.