- Time Series support – allow to store time series data and run queries on time series of any size in milliseconds.
- Documents compression – will usually reduce your disk utilization by over 50%.
- Many indexing improvements, especially with regards to indexing and querying date time data.
As I mentioned, you can upgrade to the new version right now on your own instances and you can deploy new cloud instances with RavenDB 5.0 immediately.
RavenDB 5.0 is backward compatible with RavenDB 4.2, you can shut down your RavenDB 4.2 instance, update the binaries and start RavenDB 5.0 and everything will work. The other way around will not work, mind you. Once you have run RavenDB 5.0, you cannot go back to RavenDB 4.2.
A RavenDB 5.0 instance can be part of a cluster running other RavenDB 4.2 nodes (and vice versa), which allows you to do rolling migrations and test the RavenDB 5.0 in production without committing all the way in.
An application using RavenDB 4.x client will be able to just continue working with RavenDB 5.0 server, with no change in behavior. That enables you to just switch over to the new version without needing to modify the whole stack at once.
For users running RavenDB 4.2, I’ll remind you that this is a Long Term Support (LTS) release and it is perfectly fine to remain on that edition for the next year or so.
Users on the cloud that are running RavenDB 4.2 can request an upgrade to RavenDB 5.0 via our support, but for the foreseeable future, we are going to keep users on RavenDB 4.2 unless they ask to be upgraded to RavenDB 5.0.
I’m really happy that we go to this milestone and I wanted to take this opportunity to congratulate the team behind RavenDB 5.0 for an excellent job done. Under non trivial circumstances, we have a pretty amazing product shipping, and I am very proud of what we have shipped.