The new Mobot is a conversational AI assistant that helps teams explore and understand their data using natural language. We’re now extending those same capabilities to one of the most time-consuming tasks in observability: creating and configuring monitors.
Setting up a monitor by hand means writing a precise log search query, choosing the right detection method, tuning thresholds, and configuring notification routing. It also means calculating trigger windows that actually make sense for your service, which can be the difference between a monitor that catches real problems and one that just adds noise. This process can be time-consuming and particularly difficult for those who are new to their organization or to using Sumo Logic, which is why we made monitor creation conversational.
What’s new
Here are two key capabilities of Conversational Monitors:
- Create: Describe what you want to monitor and Mobot handles the rest, generating the query, recommending a detection method, configuring thresholds, and preparing a monitor for review.
- Configure: Rather than relying on guesswork, Mobot analyzes historical data to recommend trigger windows, alert conditions, notification routing, and standardized metadata.
Conversational monitors support static, anomaly, and outlier detection for log-based monitors. They also keep humans in control at every step. The same role-based access controls (RBAC) that govern monitor creation, editing, and viewing apply to Mobot, and every monitor Mobot drafts requires your confirmation before it publishes.
Example: Creating a checkout latency monitor
Context: An engineer is shipping a new checkout service and wants alerting in place before it hits production, but doesn’t want to spend an afternoon hand-tuning thresholds.
Starting prompt: “Watch my checkout latency and alert me if it looks abnormal.”
How Mobot tackles it: Mobot identifies the checkout service’s latency field, evaluates historical variance, and recommends anomaly-based detection with a rolling five-minute window. It proposes a monitor named and tagged consistently with the team’s existing alert structure, and suggests routing critical breaches to the on-call PagerDuty channel. The engineer types “change threshold to 10%” to tighten it, then confirms.
The result: A production-ready monitor in three conversational turns instead of a manual build that requires knowing the right query syntax and routing variables in advance.
Why it matters
Creating monitors conversationally with Mobot eliminates the learning curve. Instead of learning Sumo Logic’s query syntax and the nuances of your environment, you can describe what you want to watch in natural language and have a working monitor in minutes — no matter your experience level.
Mobot also takes on the work of suggesting the right alert type, thresholds, and notification routing based on your actual data, so you’re not guessing at what “normal” looks like for a new service. That adds up to real time back: an afternoon of manual monitor configuration becomes a few conversational turns.
Try it for yourself
Conversational monitor creation is available today for most Sumo Logic customers. Check out the help docs for instructions on how to get started.
Not yet a customer? Schedule a demo with our team.
