The change point embeddable utilizes the change point aggregation to fetch a change point for requested metric. If the split field is configured as well, it scans all partitions to locate change points. Once the results from the change point aggregation are obtained, it plots the metrics with the Lens embeddable, adding an annotation layer with a corresponding change point. If no change points have been found, it renders an empty state.
But for the following configurations:
No split field is selected (single metric mode), e.g. avg(responseTime)
Partition values of the split field are provided, e.g. avg(responseTime), split by hostname, partitions=['en.wikipedia.org', 'https://github.com/']
Instead of displaying an empty state, it is more informative to render the metric along with a notification indicating that a change point couldn't be determined. This ensures that users receive some form of visualization even when a change point isn't found.
This behaviour should be configurable with an additional property to the embebeddable input, e.g. allowNoResults.
The change point embeddable utilizes the change point aggregation to fetch a change point for requested metric. If the split field is configured as well, it scans all partitions to locate change points. Once the results from the change point aggregation are obtained, it plots the metrics with the Lens embeddable, adding an annotation layer with a corresponding change point. If no change points have been found, it renders an empty state.
But for the following configurations:
avg(responseTime)
avg(responseTime), split by hostname, partitions=['en.wikipedia.org', 'https://github.com/']
Instead of displaying an empty state, it is more informative to render the metric along with a notification indicating that a change point couldn't be determined. This ensures that users receive some form of visualization even when a change point isn't found.
This behaviour should be configurable with an additional property to the embebeddable input, e.g.
allowNoResults
.