Causely, a provider of an observability platform based on causal artificial intelligence (AI) models, today revealed it is now adding support for open source Grafana dashboards.
Announced at the GrafanaCon 2025 conference, this integration provides DevOps teams with an ability to surface root cause analysis within dashboards that are already widely used by many IT operations teams to visualize workflows.
Causely CTO Shmuel Kliger said that while the company will continue to provide its own graphical user interface, Grafana dashboards in many IT organizations are already a de facto standard. Additionally, Causely now supports Grafana Alertmanager to enrich existing alerts with continuously updated root cause analysis in real time.
The Causely platform is based on causal AI models that make use of machine learning algorithms to identify anomalies that it uses to infer the root cause of IT issues. The Causely system works by automatically mapping the topology and service dependencies of an application. That approach reduces manual troubleshooting by making sense of patterns to surface more actionable intelligence, said Kliger. Rather than simply identifying something might be wrong, that approach makes it feasible to identify the root cause of the issue, he added.
As AI is more deeply embedded within IT workflows, multiple types are now increasingly being used to automate workflows. Causal AI models make use of machine learning algorithms in much the same way predictive AI models do. Generative AI models, on the other hand, analyze massive amounts of data to, for example, provide summaries of events in natural language.
Each DevOps team will need to decide what mix of these AI models to employ, but while generative AI models are garnering the most amount of attention, it is causal AI models that will have the most impact on improving IT incident management by streamlining the number of alerts that would otherwise be sent when all the alerts are related to the same issue, said Kliger.
Hopefully, IT incident management will become much less stressful in the age of AI than it is today. It’s not uncommon for IT organizations to have to mobilize multiple teams just to discover the root cause of an issue that can be fixed in a matter of minutes. The more time it takes to discover that issue, the more likely it becomes that IT team members will burnout as the number of incidents that require a response mount.
At this point, DevOps teams would be well advised to review their existing incident management workflows to determine which tasks might soon be automated using a variety of AI models. The challenge and the opportunity now is determining which AI models to use when depending on the goals defined.
In the meantime, there may come a day soon when the number of incidents that IT teams need to respond to should steadily decline as more of the focus shifts toward preventing them from ever happening in the first place. After all, the best IT incident is the one that never ever actually occurred.
