Gremlins during threading

I just ran a 100+ pcs rigid tap job on the lathe and encountered some odd pauses before doing a G 33.1.
First 35 or so pcs had no issues. Then an occasional "hang" before threading of a few seconds followed by a few parts without a hiccup.
Machine would continue doing this randomly with some pauses approaching 2 min. It was painful to watch taking this long to cycle the part.
There was no tap breakage. Everything stayed in sync. Parts measured up.
I could re-load the program and everything would run as normal then it would repeat with random pauses at G 33.1.

Computer is old, also noticed a few the keyboard numbers not working.
I'm running older linuxcnc ver 2.7.7 on mesa hardware. I did jiggle encoder wiring connections in the hopes something was loose and it was not.
Tried running the scope while the job was running and that seemed buggy. Couldn't clearly read in the index. Will have to explore that again. I get some minor graphics trails from the cursor at times. Not sure if that needs addressing?

I know the spindle at speed signal is true. I assume A and B channel is working correctly or I would have broken a tap?
Are there any other index related signals I should be looking at?
Any known bugs with treading on 2.7.7?
I did encounter a similar thing a while back while doing a threading cycle, a random short pause waiting to sync up but not as noticeable as this go around.
I need to fix this.

Sounds like a missing index from the encoder
(if you are completely sure spindle-at-speed is always true)

Thanks for the responses. I'll look closer at index enable.
One other thing worth mentioning is that my wiring is going from the spindle encoder to my VFD which has an encoder card fitted and is split out in a Y to the mesa card. Shield drain terminates on one end at star ground. Wonder if the Y is possibly introducing a signal drop?
I'll have to poke around with a scope.