The max allowed distance between data points on all dashboard charts is 90 seconds. This means that if two data points are farther apart than 90 seconds, a gap will be shown instead of drawing a line between the data points. This is because if we connect the points it can be misleading -- it may appear that the points on the line in between had values when in reality they didn't.
The reason this tends to show up more often for the memory counters is that they are only sampled every 60 seconds by default, while most other counters are collected every 10 or 20 seconds. So while we can miss several samples of other counters without showing a gap, if we only miss a single collection of the memory counters, the pre- and post- data points will be 120 seconds apart, and thus a gap will be shown.
Ultimately if the SQL Sentry Monitoring Service was running and we miss a sample, it means that load on the target and/or the network was too high, or some other problem prevented it from collecting counter data. In other words, it's indicative of a problem on the network or monitored target that needs to be investigated. Typically you will also receive some failsafe alerts (Error or Offline) when this happens, assuming those alerts are configured. It could be something as simple as the network went down temporarily, or the target server was rebooted.
You can often diagnose the root cause by looking at what was happening on the other charts immediately prior to the gap, since again most of the other counters are collected more frequently than memory. Or click the “Jump to Calendar” button so you can see the chronology of events during that time.
If you still aren’t sure what you are seeing, run the “Windows + SQL Server Performance” report for the range in question (be sure the range size is <=30 minutes so we have raw data), export it to Excel format and email it to us at email@example.com and we’ll take a look.