There are two answers to your question on who keeps the records, the first is equipment manufacturers, who may or may not share the data for liability reasons.
The second,and really who I was talking about, is any agency or group promoting a specific gear requirement or configuration. Right now that is mostly the DIR agencies, those are the groups promoting the “failure point “ discussion” By and large they expect folks to use their directed equipment configuration but do not give any specific data to support the likelihood of failures in their failure points.
As a specific example. If someone is going to say you must use stainless steel cam bands or no plastic buckles I would expect you to provide some data that shows actual failure rates of plastic cam bands or buckles to support the requirement, not just anecdotal evidence that They once saw a cam band break. Alternatively if there is a higher risk of breakage a solution may be periodic replacement of that item. Another specific example might be requiring the use of spring heel straps in your fins vs bungee cord ones because the bungee cord is a failure point. Is there any systemic data to indicate that bungee cord heel straps break more often? Could a viable solution be that you must replace the bungee cord heel strap annually?
There is far too much dogma in some areas of diving rather than data driven risk management. What I am advocating is moving more and more towards data driven risk management models.