Disk fragmentation metrics

[AntonioSousa]

Disk fragmentation was a thing some time ago, but with the advent of non HDD storage, it seems to have disappeared as an issue.

We have clients where HDDs are still important on some servers though. And I have been asked for fragmentation counters.

Now, I know that Dynatrace doesn't have any of these metrics, but I believe there are other indicators that are related, including especially disk latency, and Queue length. While not directly related, inode availability is also an interesting indicator on Linux, where fragmentation also exists, and a high level of inodes might mean a lot of small files.

Has anybody had issues lately with this topic?

[ChadTurner]

interesting, cant say I've had this topic come up at my organizations. Did you come up with any solution @AntonioSousa 

[AntonioSousa · Accepted solution]

@ChadTurner, not yet. We have focused on disk latency instead. In my opinion it's a similar metric.

[Julius_Loman]

Fragmentation is not something you can measure continuously. So I'd recommend looking the disk latency.