This message was deleted.

The aggregation interval is independent of the scrape interval. The current minimum interval is 60s and is dynamically adjusted based on the time range for long intervals.

@Srikanth Chekuri can you please point to the part of code where one would need to make adjustments in order to make the minimal interval go down to 1s in a self-hosted version? Thanks much!

Curious why do you want to do that? The step interval is auto-adjusted based on the time range selected like how Prometheus does and recommends.

Yep, adjusting it dynamically makes sense but the lower bound of 60s is too high for some of our metrics. There is a set of them where the granularity of at least 1s is required for us to get an adequate picture of the system's behavior.

I see, that's a valid point. This is the relevant frontend code that decides the step interval https://github.com/SigNoz/signoz/blob/0c4149225f8a338197bdd2c68981bd1e0c2a330d/frontend/src/lib/getStep.ts#L38