Excellent question. The performance policies you've selected for your SQL logs and SQL data are designed for volumes directly connected to a Windows host. There is one attribute on both of these policies that makes them less ideal for VMFS Datastores. This attribute, called "Quota Exceeded Behavior", is set to "Set Offline" in both of these policies. This setting will cause a thin provisioned volume to go offline in an out of space condition. ESX will respond to this behavior by going into APD (all paths down) and potentially PDL (permanent device loss). The performance policy named "VMware ESX 5" has the correct setting, which is "Set to Non-Writeable". ESX will respond to this behavior by "stunning" (pausing) the VM. ESX is able to automatically recover from this situation once the low space condition is resolved.
To address this, you might create a new performance policy for both MS SQL and MS SQL Log on VMFS. You can use the existing policies as models. The only change would be to the "Quota Exceeded Behavior" attribute, which should be set to "Set to Non-Writeable". Because the block size and the application category will remain the same, you should be able to change the performance of the volume.
I hope this helps,