Modern drives have embedded monitoring. Whilst the disk hasn't failed, something internally is telling the embedded monitoring that it may soon fail - it's running a little hot, having excessive write errors, latency is drifting up... something subtle is going wrong.
It could be caused by a firmware error... but I'd take it at face value.
It might have failed by the time you read this, or it might work fine for another 20 years.
Either way - when I was a server admin I would get any disk with a predictive failure swapped out immediately. No point waiting for it to fail - even on a RAID system when a disk fails although you don't lose data you have your trousers down as a coincidental second failure would ruin your day!
Please remember to rate helpful posts to identify useful responses, and mark 'Answered' if appropriate!