How AI Makes SLOs Easier to Define and Problems Easier to Resolve

We are building AI into Nobl9 so the platform carries more of the SLO workload, and this session is the first public look at where that is headed. The two capabilities we will show sit at opposite ends of the SLO lifecycle: deciding what to measure when a service has no SLOs yet, and working out why an error budget is burning once it does.

SLO discovery is a conversational AI assistant in Nobl9 Labs. Describe a service in plain language and it interviews you about user journeys, failure modes, and what acceptable quality looks like. It then drafts SLO proposals for your team to review, each with a recommended SLI, a suggested target with written rationale, and the calibration questions to answer once real data is flowing.

The alert explainer is AI for the moment a budget starts burning. Nobl9 knows the SLO is burning, and the cause lives somewhere in your infrastructure. The explainer connects to your own observability stack, investigates the alert against your traces and metrics, and writes its findings into Nobl9 as an annotation. In testing it has traced a burn to a single oversized response and flagged alerts that were simply noise.