As XtremIO becomes more popular with customers, there will be a need to integrate the statistics it provides into third party monitoring systems. XtremIO has a very simple UI to work with and it shows you a decent amount of statistics already in the UI, per volume; IOPs, latency, bandwidth, etc… EMC already has a vCenter Operations adapter for the VNX and VMAX called EMC Storage Analytics (ESA). ESA is a product you can purchase and it integrates very easily into vCenter Operations. The plugin lets you pull in a TON of resources from a VNX/VMAX array into vCenter Operations. Once that data gets ingested into vCenter Operations the powerful analytics engine, the life blood of the product, can be leveraged against your storage resources. Now you can analyze an issue, perform root cause analysis or do capacity planning from the application layer all the way down to the storage, that’s powerful .
From my understanding, XtremIO is on the ESA roadmap . That means, sometime in the relative near future, XtremIO customers who have the advanced version (or better) of vCenter Operations can use the ESA to bring their x-bricks into vCenter Operations. If you were at VMworld, or EMC World (don’t remember which or if both), the Hands on Labs were showing the awesome sauce of XtremIO performance metrics inside of vCenter Operations. So, if it’s only on the roadmap for ESA, how is it being shown already? Enter Matt Cowger. Matt works for EMC as a…well, I’m not sure what his title is, but he’s really smart. Matt wrote some PoC code in python to demo this capability. The code connects to XtremeIO via the XMS to get all the bricks and volumes connected through that XMS, along with their corresponding metrics. The results are returned in JSON and then those metrics are parsed and pushed to vCenter Operations using the HTTP POST adapter. I wrote a little bit about the HTTP POST adapter for vCenter Operations here. I took his script and put it in the Varrow labs and started pushing XtremeIO performance metrics into vCenter Operations. Below is a screenshot of a few metrics I put on a metric graph.
- Top metric is the write IOPs for a particular volume. The high is 4,613
- The middle metric is the read IOPs for that same volume. The high is 6,792
- The bottom metric is the write latency for that same volume. The high is 974. Keep in mind that metric is measured in microseconds. So the highest that volume has experiences is less than 1ms. The average is much lower than that
Once XtremIO is fully integrated into ESA I imagine it will come with lots of pre-built dashboards so you don’t have to mess around with making your own
This post is really meant as just a preview of what some metrics from XtremIO might look like inside of vCenter Operations. Big shout out to Cowger for writing the code to show the possibilities. I’m looking forward to see how XtremIO gets implemented into ESA and what the pre-built dashboards are going to look like.