tl;dr> Today we've released v0.6.0 of GlareDB, featuring the very first iteration of the GlareDB Pro offering, along with a host of other features and fixes including PRQL support and use of S3, GCS, R3 and MinIO in Python and the local CLI.

GlareDB Pro: Simple and Transparent Pricing

As part of the v0.6.0 release, GlareDB Cloud now has a Pro offering which we're really excited about and we think you might be too!

Not excited about billing?

Here's the thing: compute resources aren't free and we've all seen cloud platforms that are complicated and even a bit hostile to users. We laid out several goals to ensure that our model is simple and transparent:

• GlareDB Cloud is a serverless platform, and the model must reflect that. It has been built and designed from the ground up with serverless usage in mind.

• You should only pay for exactly what you use. If you don't use any resources then you aren't billed for anything.

• GlareDB's goals have to be aligned with you, the user. If we charged you for the amount of time it took to run a query, then you wouldn't have as much transparency and control on cost. Furthermore, we wouldn't be inclined to optimize performance: if we make it faster, your bill goes down; if we provision more resources, your bill goes down and it costs more.Thus we chose to base our billing model on metrics that we can record and provide transparency for.

• The pricing structure has to be easy to understand. Too many billing systems have pages and pages to explain all of the fees and different components which have different rates, grades and tiers.

GlareDB Pro and GlareDB Free Plans

Based on these principles, we introduce the GlareDB Pro plan:

• Unlimited data scans and writes: $7 USD per TB, billed monthly
• Unlimited data storage: $23 USD per TB, billed monthly
• Fully serverless compute
• Up to 5 deployments (logical GlareDB instances)
• Up to 30 organization members

By comparison, the GlareDB Free plan is now:

• 256 GB of data processed per month
• 10 GB of storage (total)
• Fully serverless compute (technical preview)
• 1 deployment (logical GlareDB instance)
• Up to 5 organization members

Serverless compute is used when connecting to a deployment via direct connection and in Hybrid Execution using Python or the CLI.

That's it. Simple, transparent and no hidden fees or complex criteria.

How it Works

We're pretty proud of the implementation. We aggregate bytes processed by all queries. When the end of the billing period is reached, the bytes processed are aggregated and multiplied by the cost ($7 per TB), rounded to the nearest cent. There are no other forms of rounding, and no floating point math: the bill really does reflect the usage.

Bytes processed is the sum of all bytes read out of original storage and into the GlareDB server by the query along with the size of the payload written either to a table or returned over the network to the user. Many storage formats, like Apache Parquet, make it possible to minimize the amount of data read out of storage; while some, like CSV, require reading more data than others. In any case, it's possible to write queries and store data in ways that can limit the amount of data read: you have control over your bill, and our improvements impact latency and the operational overhead for us.

GlareDB Pro will evolve with GlareDB and in response to your feedback and needs. GlareDB Cloud will always strive to be the best serverless platform by being aligned with your needs and easy to understand.

Additional features

In addition to GlareDB Pro, there are host of other amazing features in v0.6.0:

• Preliminary PRQL support, which is an analytics-focused query language built around the concept of pipelines (announcement coming soon)

• Full support for GCS, S3, R2 and MinIO providers in Python and local CLI

• GlareDB is improving support for direct INSERTs into PostgreSQL and MySQL data sources. See epic 1143 on GitHub for the complete roadmap for all data sources.

See the full 0.6.0 release notes for more details.