Can it be used to debug system problems as well?
It can assist with communication problems, mostly by allowing quick isolation of when an issue occurred to speed the finding of a fault. Patterns will develop as normal activity and thus abnormal activity will stand out as not fitting a visual pattern.
One example of system diagnosis is lack of resourcing for IVR ports. Typical behavior would have a short yellow (ringing) block followed by a blue (established) block as a message is played. Long yellow blocks with no blue block following indicate continuous ringing and a sign that not enough resources are available to play the desired message.
Another example is round robin configured DN selection from a bank of DNs. There is an evident cascading pattern that should occur in that scenario and deviations from that pattern may point out issues with that bank of DNs.
Another example is default routing. The cessation of activity on a bank of DNs to then only see activity on a default point would indicate a system failure.
Queries can also be used to display errors as captured from the TLib event stream.