Obsolete data is excluded from query results by default, this can have an impact on the way data appears once you apply advanced functions to your query.
Aggregation functions like sum() or avg(), ect will group time series together, however if some of the time series stopped reporting in more than 14-days ago they are not included in the aggregated results.
Enable the “Include Obsolete Metrics” property on the chart to visualize all time series which have been classified as obsolete.
Using functions that aggregating data, the resulting aggregate values will not include the obsolete time series and can lead to inaccurate results.
The screenshot below shows an average of request latency without the "include obsolete metrics" property being enabled on the chart:
In the example above, the aggregate values range between .01 and .03 because it's only aggregating on the subset of test data that is not marked as obsolete.
When we enable "Include Obsolete Metrics" in the following chart, then the query results are based on our entire test data set. Now the range of aggregate values is between 1 and 3.
When looking at historical data, it can be tricky to identify if all data is being included in your query results. In this case, we would suggest enabling the “Include Obsolete Metrics” property to ensure all data is included in your query results.