Your VMware vRealize Operations (vROps) environment is up and running and collecting vSphere objects and metrics from vCenter. It might even be collecting non-vSphere objects and metrics from other targets via vRealize True Visibility Suite (vRTVS) management packs.
I personally have a single node cluster collecting data from vCenter and many other targets, it looks like this in Administration - Management - Cluster Management.
vROps clusters are sized based on how many objects and metrics you're collecting and how long you're keeping the data. I use the following URLs for vROps sizing exercises:
https://kb.vmware.com/s/article/2093783 - KB listing all versions of vROps with links to their respective sizing guidelines.
https://vropssizer.vmware.com/sizing-wizard/choose-installation - interactive tool used to provide sizing recommendations.
Built based on these recommendations, each node in your vROps cluster is given CPU, Memory, and Disk. CPU and Memory are allocated based on how many objects and metrics you're collecting. Disk is provisioned based on the same, but also how long you're keeping the data.
Adding and removing CPU/Memory is fairly straight forward. It's often times done when additional vCenters (or other targets) are added to vROps, in turn capturing additional objects and metrics.
Adding and removing Disk is also relatively routine, but what happens when you run out of disk space before expanding? And how do you clean up disk space if you'd prefer not to allocate more?
First, what does disk look like on your vROps cluster? Each node will have five filesystems:
/ - root fileystem
/boot - boot filesystem
/storage/core - support filesystem, support logs are stored here.
/storage/db - database filesystem, data (objects, metrics, properties, historical data) is stored here.
/storage/log - log filesystem, logs are stored here.
When you run out of space it will most frequently be in /storage/db, where your data is stored. When this happens, vROps will crash and more disk will need to be provisioned to the filesystem. There are times you've already allocated more than you want and you'd like to clean up these filesystems. While it's easy to remove old logs from /storage/core and /storage/log, it's not as easy to cleanup /storage/db. In fact, it's not recommended at all, so what do we do?
We use the Global Settings in Administration - Management. vROps performs self-maintenance nightly (still haven't found the exact time) based on the settings configured here. These are the defaults.
Based on these settings, my vROps master node disk consumption looks like this. Note: this VMware VM Summary dashboard is available out on VMware {code}.
Notice that /storage/db is 205.7GB in size, with 97.34GB used, or 47.32% used. I'd like to buy back some of this space, so I don't have to allocate more. Let's adjust some of the Global Settings to do this. Specifically, I'm going to adjust the following:
Action History - number of days to retain the historical record of remediation actions. Adjust from 30 days to 7 days.
Deleted Objects - number of hours to retain objects that no longer exist. Let's adjust from 168 hours to 24 hours.
Object History - number of days to retain the object configuration history. Adjust from 90 days to 10 days.
Symptoms/Alerts - number of days to retain the canceled alerts and symptoms after the cancellation. Adjust from 45 days to 30 days.
Time Series Data Retention - number of months to retain statistical sample data for objects. This is likely where you will save the most space. Let's adjust from 6 months to 3 months.
Additional Time Series Retention - number of additional months of rolled up data. You will likely save a lot of space here. Let's adjust from 36 months to 0 months.
Deleted Users - number of days to retain users and their content after they are archived. Adjust from 100 days to 30 days.
Our Global Settings now looks like this.
Let the cleanup run over night and check back on your filesystems the next day. This is what mine look like.
Notice that filesystem /storage/db on vr80-bpeterson went from 47.32% used (97.34GB) to 28.52% used (58.66GB), I bought almost 40GB of space!
This is a nice alternative to simply adding more disk to your vROps cluster. Turn the dials on Global Settings to meet your needs!