This post is to report/complain about/request a fix for the inconsistent units used by the memory stats in the GUI.
Cluster Summary: GiB (base 2 / 1024)
Node or VM Summary bar (in the first section with the hostname): GiB (base 2 / 1024)
Node or VM Summary Memory usage graph: GB (base 10 / 1000)
For example, I have a node with 62.7GiB of RAM showing in the top section, whereas on the Memory usage graph it shows as 67.33. 62.7*1024*1024*1024/1000/1000/1000 = 67.32, so this is clearly due to the difference in units/conversion used. This affects both the "used" memory as well as the Total.
Considering that GiB (Mib, KiB, etc) is what is reported generally, e.g. by "free -hm", and is also the units used when allocating memory to a VM, it makes sense that the Memory usage graph should also use these units.
Cluster Summary: GiB (base 2 / 1024)
Node or VM Summary bar (in the first section with the hostname): GiB (base 2 / 1024)
Node or VM Summary Memory usage graph: GB (base 10 / 1000)
For example, I have a node with 62.7GiB of RAM showing in the top section, whereas on the Memory usage graph it shows as 67.33. 62.7*1024*1024*1024/1000/1000/1000 = 67.32, so this is clearly due to the difference in units/conversion used. This affects both the "used" memory as well as the Total.
Considering that GiB (Mib, KiB, etc) is what is reported generally, e.g. by "free -hm", and is also the units used when allocating memory to a VM, it makes sense that the Memory usage graph should also use these units.