This is why JPI EGT probes fail so often (I machined this one off at a slight angle in a milling machine):
They form the thermocouple (Type K) junction by welding the two ends at the tip of the probe, so you need just 0.2mm (I measured it under a microscope) of the stainless steel to wear away before the tip gets eaten.
This may get you a fast response time but is this needed? Their competitor EI reckons not and they claim their probes are more robust. I also don’t think you need it because the time constant in a gas flow of that sort of velocity will deliver a TC of the order of seconds even in a 1mm thick 4mm diameter stainless steel tube. OTOH the encapsulant which is some sort of ceramic (obviously it can’t be epoxy 
) might slow things down.
But anyway if you are going to construct an EGT probe in this way, use more than 0.2mm of metal…
BTW the above pic was taken directly with a Samsung S10e phone!
