How Should I Aggregate Data?

Because Observe can support many different use cases with many types of data, it can sometimes be difficult to select the correct OPAL tool. Here’s a way to pick the correct aggregation term for the task at hand:

• If you need to align metrics by time, use align. Alignment places time series data into regularly spaced bins on a grid so it can be compared with other aligned time series.

• If you need to aggregate numbers by space, use aggregate. Aggregation computes a single numeric value for each of the points in an aligned series so that series can be compared. If you look at visualizations automatically created by the Metrics Explorer, you will often see align and aggregate used together.

• If you need to align any type of data across time and space across the whole query window, use statsby. The statsby command aligns and aggregates in a single command. It is not accelerable.

• If you need to align any type of data across time and space, either bucketized “for every X minutes” or across the whole query window, use timechart. The timechart command aligns and aggregates in a single command.