Introducing Natural Language Queries: Turning plain English into insight

The syntax tax of complex query languages like PromQL creates a steep learning curve that bottlenecks incident response. This friction slows down onboarding for new engineers and hinders those who only use observability tools infrequently. Chronosphere’s natural language queries eliminate this barrier. By translating plain English – as well as other languages – directly into valid PromQL and log syntax, anyone on the engineering team can easily query telemetry data, bypass syntax errors, and get immediate answers.

PromQL offers deep capabilities for metrics analysis, but its complexity frequently acts as a barrier to the underlying data. Even seasoned developers can struggle to recall the exact syntax for a histogram_quantile or how to properly format multiple label filters under pressure. This forces engineers to constantly switch contexts between the active investigation and query documentation. During a high-stakes incident, every second spent fighting syntax is time lost on actual resolution.

Natural language query generation solves this problem by taking your operational intent and instantly translating it into a valid PromQL statement. Instead of relying on trial-and-error debugging to build a query from scratch, you get an immediate, accurate starting point for complex investigations. This keeps your focus strictly on diagnosing system health and resolving the issue rather than wrestling with the query language itself.

The difficulty of query syntax isn’t limited to metrics; log analysis frequently suffers from the same overhead. Finding the right log lines across different services often requires knowing specific search operators or field names that vary from one platform to another. This creates a “blank page” problem where an engineer knows the specific data they are looking for but is forced to spend time figuring out how the tool requires them to ask for it.

Natural language queries remove this friction by providing a consistent experience across telemetry types. You can simply ask the platform to filter for specific error codes or aggregate by metadata tags using plain English. This consistency ensures an uninterrupted workflow, allowing you to maintain your train of thought and find the root cause faster as you move from metrics to logs.

Rarely is a query perfect on the first attempt. Investigations evolve as new data surfaces, often requiring engineers to pivot their perspective or drill deeper into a specific subset of data. Typically, this means manually editing strings, adding brackets, or looking up regex patterns to narrow down a search.

The refinement capability allows for a conversational approach to data exploration. Once a base query is generated, it can be modified by simply providing further instructions in plain English—such as “now group this by region” or “only show results with a 500 status code.” This iterative process keeps the engineer in the flow of the investigation, allowing them to follow a trail of evidence without being interrupted by the mechanical requirements of the query language.

The days of debugging your own queries during an incident are over. With Chronosphere’s natural language queries, you can focus on resolving issues based on the data, not your memorization of syntax.

• Accelerate onboarding: Allow new hires to query telemetry from day one.

• Reduce cognitive load: Let the platform generate the PromQL and log syntax for you.

• Maintain context: Move seamlessly across telemetries without changing your workflow.

Natural language queries are currently available for Chronosphere users. Reach out to your account team or book a demo to see how conversationally querying your data can speed up your incident response.