When I dissected a failed IP it had much metal in the body of pump. I believe (this is consistent w/GM Tech Bulletins) the metal came from the original metal rollers and/or IP internal wear surfaces associated w/the rollers.
I.e. the metal did not 'cause' the failure, but is the failure. The higher the rate of metal shaving production, the faster the IP fails. I'm thinking the metal shavings are wear of rollers or their bearing surfaces & that wear decreases reliability of the req'd ijection pressure & therefore reliability of injector opening pulses (failure = DTC 35/1216 and/or 36/1217).
\'95 P30 Motorhome 16,500gvwr, 26,500mi<br />New Ipump @ 25k, boost, EGT, homemade remote FSD cooler