I find DPA invaluable to identify and troubleshoot issues, but it would be great to see some more at-a-glance type metrics added to quickly identify how hard servers are working and identify trends.
Each of these is available, either by querying the database or in the case of virtuals by logging into Virtualization Manager (VMAN):
1) Peak and Average IOPS during the last hour and last 24 hours.
2) Peak and Average Throughput during the last hour and last 24 hours.
3) Read/Write Ratio during the last hour and last 24 hours.
4) Peak and Average TPS or TPM during the last hour and last 24 hours.
Not sure where in DPA those could be added, but those are real numbers that I have to collect manually on a regular basis for key servers. These 16 values combined with Peak CPU during the last hour/24 hours allow you to profile your servers over time, give you the numbers you need to work with Storage admins and vendors, and give you values that you can present to management.
