Recent reproducibility issues and increase in peak fronting after consecutive injections

The model number on GC is 6890N (G 1540N). I am using a method with split injection with 50:1 split ratio on hp-5 column to measure concentration of a methanol/water mixture using ipa as an internal standard. The method has been successfully used giving a reasonable standard deviation for several months, and recently the standard deviation increased 3x. I tried replacing the liner, septa, gold base seal, reinstalling column, and the split vent cartridge.  After replacing the liner, septa, gold base seal, and reinstalling column, I started noticing some variation in the retention times of the peaks I was looking for. It was suggested to me to change the split vent cartridge as well so I did so, now the peak response on the FID also decreases pretty dramatically on consecutive injections, and the peak fronting still occurs. 

I have also noticed something else, when running the method when the split vent valve turns on (time 0.2) the error light in the software turns red until maybe time 0.4 and the dialogue box says something about the inlet flow being too low and the column flow rate drops from 1 ml/min to 0.9 ml/min, and then turns green again before the switch turns off. Then when the turns off again the error light turns red again shortly after the valve turns off and then turns green again.  Is this normal for the GC or is this another symptom of the problem?

My next step is to call service, if anyone has any ideas on what might cause this sudden loss of performance please let me know.  And if you need more information let me know.