Time-series aggregation step in diff / comparison queries

[kolesnikovae]

The time series always uses sum aggregation, and the aggregation window depends on the time range.

For example, when you query the last hour, each point represents the sum of values within a 5-second window. However, when you query the last 12 hours, the window increases to 30 seconds, so each point covers a larger time span and aggregates more data. This results in higher sums for each point, which can be misleading if you're expecting consistent, time-normalized values across different time ranges.

That said, I find this behavior quite confusing. I believe users expect to see series where the values are normalized across different time windows. This inconsistency feels like a bug, and I think we should fix it to provide more intuitive results.

In the screenshot below you can see that in the diff view, the difference is even more pronounced and confusing: