We're migrating out of one of our older data centers currently, and sending them to either Azure or one of our two VM data centers. We've been working on getting all the servers out of the old data center for a year. Being able to migrate those servers, with the help of the service teams, is one of the best features. Instead of having to do them one at a time, we can build a VPG. That is especially helpful with some of these really large VPGs. We did one a couple of months ago that was 36 terabytes. We were able to migrate that entire VPG at one time, watch it replicate once it was there, and then do the reverse replication. That process has been amazing.
We use Zerto to protect VMs and our RPOs are very solid. The RPO is a little slower for Azure, but that was expected and it was covered in the documentation that Zerto provided. But the RPOs from one VM center to another are solid. The same is true for our RTOs. We have no complaints in that regard at all.
And it will definitely help reduce downtime if we have to migrate from one data center to another due to a disaster (which we have not had to do so far). Downtime would cost us a lot, no doubt. We have not had any major disasters as of yet or problems with spyware or ransomware. But we have had instances where a server was corrupted in one data center and because it was backed up with Zerto, we were able to fail over to the secondary site and get the VPG back up very quickly. From the time that the decision was made that we needed to fail over, it took around 30 minutes. It was very quick, especially compared to trying to troubleshoot and rebuild. Our patients weren't affected.
The platform has also helped us to streamline our DR testing. We're able to do our annual test as quickly as possible, whether it's a failover rollback or a system move. It has made that process much much quicker and a lot less painful.
I've only been in this position for a year, but from what I've heard, DR testing was not a pleasant experience prior to moving to Zerto. They were always having failures and then would have to set up another downtime and test again. With a lot of our applications, those that are tier-ones, we're having to do the DR test at 1 AM or 2 AM. Nobody wants to have to do a four-hour downtime test multiple times.
And not having long downtimes for DR has helped a lot. Our customers and service teams know that we have to do these tests once a year to stay compliant. We plan for a four-hour downtime every time we do a DR test. Very rarely do we need four hours. We block out that time just in case there are issues we didn't expect but we're usually done in under two hours, including failing over, doing the testing, rolling back, and testing again. I don't know if you can actually put a number on not having downtime, in terms of the impact on the service teams, nurses, or patients.
Right now, our DR team is just me and my boss. At one point, there were three people on the team. It says a lot about Zerto when it can be used with a minimal staff for DR. Prior to moving to Zerto, the team had four members. So it was double what it is now.