Martin,
Another solution would be to have your time table set the state of another object in the model. This object could be either a dummy object or another object that was already part of the same time table as the source. In your current model you can place a decision point after the process flow source that looks at the state of the other object. If the state is "Up" then the token would continue on the previous path. If the state is "Down" then the token would be diverted to a sink.
One side effect of this solution is that the source is not delayed like it would be in the traditional way. Rather, the flowitems are just skipped. For example, if the source was going to produce a flowitem in 15 seconds when the down event occurs the traditional approach would produce a flowitem 15 seconds after the resume event. This methods would produce a flowitem 15 seconds into the down event and send it to the sink. It would then continue to produce flowitems at the previous interval (all of which would go to the sink) until the resume event. Based on the interval of the source and all the randomness of the system the next flowitem would be produced at ??? seconds after the resume event.
If that side effect won't have a significant impact on your model then this may be a more simple approach for you. If it does have a significant impact on your model then I would recommend an approach more similar to the one you gave above.
Good Luck,
Brandon
img src="https://damassets.autodesk.net/content/dam/autodesk/logos/autodesk-logo-primary-rgbblack-small_forum.png" >
Brandon Peterson
Principal Technology Consultant
>