While it's true that each tube has only been operating about 7000 hours, the fact that 15 out of 15 are still running is statistically important. Failure-rates are computed based on testing of multiple devices.
Let's suppose 1 of my 15 tubes failed after 5000 hours. The failure-rate would not be 1 failure per 5000 hours; it would be 9.7 failures per million hours based on this calculation:
1/(5000+14*7000).
Thankfully, all of the tubes are still running. But I still dont have any data telling me how long they are expected to last. The information in my Burroughs datasheet suggests a failure rate of about 4 per ten-million hours for 5031 tubes; I wish I had that info for 5092's and 6091's..
Interestingly, I have a few 5031's in my collection and some of them have partially-illuminated cathodes. They have not responded to attempts to 'clean' them with higher current.
None of my 5092s or 6091s show any sign of wearout. So, it's a mystery to me how long they will last....