No more surprises: Newest RavenDB Update

Highlights:

• RavenDB Cloud’s new Serverless tier auto-scales CPU up and down within a range you define.
• Range-based pricing helps keep cloud database costs more predictable.
• Azure Premium SSD v2 support lets you scale IOPS and throughput without buying larger disks.

Friday. 4:55 PM. And you’re about to close your laptop (your work laptop, of course, not your laptop).

But before you do, you have to log in to your cloud console and scale your cluster down. It doesn’t take long, but it’s yet another thing you still have to do yourself.

And on Sunday night, you’ll log in again to warm it back up to keep ahead of the thundering herd. Again, it doesn’t take long, but it still has to be done. By you.

But miss doing one of these things even once, and your company pays full price for 48 hours of idle capacity. Or your applications slow down to a crawl when the traffic spikes on Monday.

In both cases, management’s going to want to have a quick word with you.

The good, the not good, and the bad

Everyone loves cloud databases because they’re so easy to spin up and scale up. But what they’re not so good at doing is scaling back down.

They’re also not good at giving you a clear picture of what you’re going to pay for all that scaling before the invoice arrives.

What makes things even trickier is that nearly all NoSQL serverless databases charge per operation. So every read, every write, every storage request adds to your next bill. That’s fine when you’re small, but it’s bad when you’re growing because you have no idea what your next bill is going to be like.

So you don’t really have a choice but to keep doing the same old scale-down, scale-up routine, even though that’s not a real solution.

And the other problem

If you’re running a NoSQL cloud DB on Azure Premium v1 storage, your disk size puts harsh limits on what you can do.

Want more speed for your indices? You need a bigger disk. Want to support more regions for higher availability? Same answer. And so on.

The problem is, once you level up your disk size, you have to live with the ongoing costs forever.

So even if you have a relatively small database, say, 100GB, you might want to purchase multi-terabyte disks so you can hit the performance targets your apps need (just to be on the safe side).

And sure, it won’t cause an immediate crisis. But it will add up over invoices.

Then, inevitably, someone will look at the bills. And you’ll get a Slack that asks, “Why are we paying for 4TB of storage for a 100GB database?”

Introducing the new Serverless tier and support for Azure Premium SSD v2

Our latest update for RavenDB Cloud is designed to fix both these problems: a new Serverless tier auto-scales your clusters up and down for you, and support for Azure Premium SSD v2.

Here’s what that means for you.

New Serverless tier that doesn’t bill you per operation

The new Serverless tier automatically adjusts CPU based on your cluster’s current load, within a range you define ahead of time. Just set the minimum and maximum instance sizes you’re comfortable with, and the database scales between them on its own.

This is pretty standard for serverless tiers. But what makes RavenDB’s different is that you’re not billed per read, write, or operation. You’re only billed for reserved capacity within your defined range.

That means you see a cost range upfront before you commit, and your final bill reflects actual CPU and RAM usage, not a per-query tally.

The new Serverless tier also comes with a hard compute cap. Instances in the production tier, for example, are capped at two cores and 4GB of RAM. So your max possible charge stays around $250/month. If your workload grows beyond that, you can migrate to a dedicated instance in one click and continue scaling from there, manually.

So if you’re building a prototype, running apps with variable or seasonal traffic, or want to forget having to remember to scale down and scale up every week, your database can now handle it for you.

Decouple performance from disk size with Azure Premium SSD v2

At the same time, we’ve also introduced support for Azure Premium SSD v2, so you can now increase IOPS and throughput whenever you need. Without buying a larger disk and being forced to pay for performance you don’t need.

So you can scale up for BFCM and scale back down for the first week of January. And your storage costs stay proportionate to what you actually need to store.

Just a heads up though: Azure Premium SSD v2 is only available in Azure regions that support it. And you’ll need to be on a data disk instead of a root-only disk. (You’ll be migrated to a data disk automatically once you hit 1TB, but you can request the migration earlier if you need it.)

How this compares to what else is out there

Most NoSQL databases offer some sort of automatic scaling or a serverless tier. So this isn’t anything new. But what is new is the way we’re going about it.

Charging per read, write, and storage operation makes sense for database providers, but it’s unpredictable for you. And the more your application does, the harder it is to estimate what you’ll owe at the end of the month.

With the Serverless tier’s range-based pricing, you’ll know the boundaries of your bill before you commit. And the hard compute cap puts a ceiling on what the database can charge you.

On the storage side, most managed database offerings still tie performance to instance size. Having Azure Premium SSD v2 support in RavenDB Cloud is a step away from that, giving you performance you can dial up or down without making a provisioning decision that’ll come back to bite you.

Who this is for

The new Serverless tier makes the most sense for:

• Devs starting a new project or building a prototype who want predictable costs from day one
  
• Teams running workloads with variable traffic, such as business-hours applications, seasonal products, or anything else that doesn’t run at full load 24/7
  
• Existing RavenDB Cloud customers who are currently managing costs through manual scaling (and would rather not)
  

One small caveat: if you’re migrating a large on-prem deployment to the cloud, the Serverless tier may not be the right fit. That’s worth having a conversation with us before you commit.

Also, Azure Premium SSD v2 support is specifically for existing RavenDB Cloud customers who are already on Azure. So if you’re considering upgrading your disk to hit performance targets but are worried about buyer’s remorse, this is something you should explore.

Why this matters

Since RavenDB was founded, our approach has always been that you shouldn’t have to babysit your infrastructure. We’re the database that optimizes itself, not the database that requires you to tinker with it, day in and day out, just to stay efficient.

These two new additions to RavenDB Cloud continue on our promise: auto-scaling that handles what you used to do manually, and storage performance that adjusts to your needs.

Even though neither of these is a major update, they can make a difference to you if you’re tired of living with the workarounds.

Check out these links for more:

Learn more about the new Serverless tier here

Learn more about support for Azure Premium SSD v2 here

Want to try RavenDB Cloud? You can start for free here.