Granularity in Load Test Results and Chart Display
Granularity is the time window that each point on a result chart summarises. Choose Raw to plot every sample, or a wider window (100ms, 500ms, 1s, 1min or 5min) to plot one averaged point per window.
Understanding Granularity
The Granularity buttons sit above the charts on the Overview, Timeline and Errors tabs of a result. The default is 1s. A narrow window shows fine detail, a wide window smooths the line and loads faster.
Raw, 100ms and 500ms are available on paid plans. On the free plan you can use 1s, 1min and 5min.
The Impact on Charts
When we translate the concept of granularity to charts representing load test results, it dictates how data points are displayed:
High Granularity (Fine Detail): If the granularity is high, you get a detailed view. This is the Raw button. Each request received during the test is represented individually.
Low Granularity (Averaged Detail): On the other hand, if you're viewing averaged values, the data points on the charts are consolidated. So, instead of seeing each individual request, you might see average values over intervals like 100ms, 500ms, 1 second, 1 minute, or 5 minutes. This offers a more general view, useful for spotting overall trends or patterns over the test's duration.
Why Does Granularity Matter?
You might wonder why we need both detailed and averaged views. Here's why granularity is crucial:
Spotting Anomalies: With high granularity (like the Raw view), you can identify specific instances of anomalies. If a single request took an unusually long time, you'd spot it here.
Viewing Trends: Lower granularity, where results are averaged, helps in understanding general performance trends. It's easier to see how the system performed over time without the 'noise' of every single data point.
Periodic Issues: Averages over 1 minute can show spikes that repeat at a regular interval, which points to a periodic problem.
Sharing Results: For a quick overview or for stakeholders, an averaged view (like values consolidated every 1 minute) is easier to read.
Making the Choice
So, how do you decide which level of granularity to use? It's all about what you aim to achieve:
If you're deep-diving into performance issues or trying to pinpoint exact moments of failure, go for the Raw tab. Embrace the detail!
For general analysis, trend spotting, or sharing results with a broader audience, the averaged views, whether it's every 100ms or 5 minutes, might serve you better.
Granularity only changes how the charts are drawn, not the test itself, so you can switch between levels as often as you like.