Live versus historical data may not always be the same. There are some reasons for this:
1. They may not match due to differences in query time periods. (e.g. Different resolutions or starting and ending times)
2. Live metrics consider random metrics out of the results to show the filtered 100 while historical metrics always consider metrics with data points.
3. Top N in live mode considers the latest 15 second data to show the Top N metrics on the graph. In contrast for historical mode, the entire eight minutes or custom time zone is aggregated to show Top N metrics.
4. Cluster performance issues such as a slow or non-responsive collector. If the collector does not respond within the time limit, then the MOM proceeds with the exiting data.