VROPs has gotten a metric crap-ton better in the custom reporting and data export feature land. However, I'm still running into one road block. How do I export a CSV file of time series data for specific metrics for vSphere objects like ESX Clusters? For instance, I would like to simply get a report that shows hourly memory usage values for every cluster for the last 7 days. Sometimes, I would like to be able to export the daily memory maximum usage for the last 3 months for each cluster.
The reason for this is so that I can do my own analysis within Excel when needed for special projects. This is to address analysis requests from management that just don't fit neatly into the VROPs out-of-the-box features.
Thanks in advance for any tips. In past gigs, I build a capacity management system on top of Splunk that allowed custom slicing and dicing of any data. I'm hoping that VROPs can get me to that level of openness with their data. Hotel California capacity management systems are not ideal.
-Sean Clark, @vSeanClark
You can do it with Reporting - create one (or more) views with the relevant metrics/period etc, then add those to a Report, and select format = cvs.
another way is to do it via the SDK which is now available - if you're into programming,there's a REST API which lets you do lots of stuff
REST API sounds like my path. I've been working with the views and CSV. There doesn't appear to be a way to select or create a view that can provide what I need. Everything just provides point in time or average over the selected time period without the detail.
Thanks for your input.
no worries-
but what details do you need which are not there in a view - can you give an example?
Imagine I have dozens of clusters. I would like to produce a CSV export of historical usage data for each cluster in one export. Below is a simple example of how it might look. In reality I would like to also include vSphere tags or other information that accompanies the host, but for now lets start simple. Don't get hung up on the timestamp format, I just put something down for an example. My idea is that I have the performance metric value with each rollup interval. The case below would be for cluster memory usage for each rollup interval.
ClusterName, TimeStamp, MemoryUsage, MemoryTotal
Cluster 1, 100000005, 40000, 80000
Cluster 2, 100000005, 43000, 80000
Cluster 1, 100000010, 41000, 80000
Cluster 2, 100000010, 43005, 80000
Cluster 1, 100000015, 41000, 80000
Cluster 2, 100000015, 44000, 80000
Cluster 1, 100000020, 41000, 80000
Cluster 2, 100000020, 44500, 80000
Cluster 1, 100000025, 42000, 80000
Cluster 2, 100000025, 45000, 80000
Cluster 1, 100000030, 43000, 80000
Cluster 2, 100000030, 46000, 80000
.......
......
FYI: I'm hearing through my vGrapevine that there isn't a way to do this through the UI. vGrapevine is looking into ways to produce this data via vROPs API.