memoryPressure is a unique nworks metric- so you won't see it in vCenter client
Essentially, it tells you how over-committed you are for memory on the ESX Host.
We calculate memory pressure from various factors - the memory allocated for all powered-on VMs, memory saved by TPS (transparent page-sharing) between VMs, the memory overhead for each VM, and various other values.
memoryPressure is intended as a kind of 'capacity-planning' or pro-active metric. You can have high memory pressure, before
you have an actual performance problem. High memory pressure alerts you that you have allocated a lot of memory to VMs, relative to the physical memory in your ESX Host - you should be aware that, if those VMs all start to demand their allocation, THEN you could have a performance problem (memory ballooning, even swapping, etc).
If you have high memory pressure on one Host in your Cluster, then it indicates you should perhaps migrate VMs to get a better balance of page-sharing across Hosts. For example, placing VMs with common OS version on the same Host will maximise page-sharing efficiency.
If ALL your Hosts have high memoryPressure, then it perhaps is time to add a new Host - or to review your memory allocation and usage for your VMs. You can use the charts and reports in the MP to do this. One report many customers find useful is the 'Bottom resource users' report for VMs. This can tell you which VMs are using LEAST resources - and maybe you can lower the memory allocation for those VMs.
You can of course decide that you are happy to over-commit quite highly in your environment - in which case, you can override our MP's memoryPressure threshold in the usual way, to suit your needs.
Hope that helps - any questions please let us know!