RPO, or Recovery Point Objective, is the maximum tolerable period (time) of data loss in the event of an incident. For example, if your VMware environment has an RPO requirement of 1 hour, then in the event of an incident such as user error that accidentally deletes a virtual machine, you should be able to recover a version of the virtual machine from an hour ago at the latest.
Nimble’s efficient snapshots allow you to effectively protect your critical data by setting the snapshot schedule to match the RPO requirement for your company – if an event occurs where data needs to be recovered, you can simply revert to the most recent snapshot. The requirement itself varies depending on business needs, but it can be powerful to see what other Nimble users are using for their RPO requirement to better inform your planning.
The graph below shows real RPO data across the Nimble customer base (taken on January 18th) for volumes with vSphere performance policies, bucketed into time periods.
The majority of RPOs for vSphere, over 75%, are currently in the 15 minute to 1 day range, where virtual machines are likely running business critical applications. The most common bucket at just over 35% is in the 6hr – 1 day range. Some businesses even have RPOs of 5-15 minutes.
Interestingly, almost 25% of volumes are bucketed as ‘None’, which means those users will not be able to revert to a previous snapshot to recover lost data. It could well be that those volumes are not production volumes, and don’t need any data protection. But if you fall into that last bucket and it’s a surprise, you should check that your snapshot schedule is set correctly! And Nimble can take it a step further and help you really sleep well at night - you can set an InfoSight alert to open a support case and warn you whenever a volume no longer has snapshot protection, so you never have to fall into that last bucket again.
So, what’s your RPO for vSphere? Should it be higher or lower? How does it vary by application?