Cube Reviews & Product Details

I find Cube to be a very sophisticated product that allows me to design complex metrics and dimensions in a scalable and composable way. I also appreciate the useful caching (preaggregation) layer. Having a strong common semantic layer is key for us as a data product company serving dashboards and insights to our customers. Review collected by and hosted on G2.com.

The service provided by the cloud team is horrible. We are a paying customer, with an annual commitment, and are having big issues when scaling that are not being solved by them. We don't have visibility into what is happening. Review collected by and hosted on G2.com.

Cube is a platform that presents the data in a very clear way. When we do the selection of column, it can also automatically generate sql code for SSMS. The query history is also great for tracking the error. When there's any problem, cube dev team is super easy to access, and they respond very fast. They really provide a strong tech support. Review collected by and hosted on G2.com.

Sometimes they query history loaded longer than expected. Review collected by and hosted on G2.com.

I'm using CubeJS for a lot of data apps on my company and it is really incredible for leveraging data for my applications. I also use CubeJS daily for manual analytics, like taking a look at the charts or copying some data to perform some stronger statistical analysis.

The interface simplicity of consumption for REST APIs is really powerful for integrating in any language and maintaining a standardized consumption around the tech ecosystem.

This can also be used as a "Feature Store" of some sorts, together with the preAggregations layer, for ML and AI applications, making it really really simple to roll them out.

I've already used the Customer Support and their response was really fast and the person helping me was also really proactive.

The ability to test changes to a cube or creating a cube via Git branches is **really powerful**! It really helps!

Finally, implementing cubes, measures, and also configuring dimensions is way too simple: we use it to connect to BigQuery and maintaining the semantics of metrics on CubeJS, which makes it so easy to understand data. Review collected by and hosted on G2.com.

I still think documentation is a bit confusing for some features.

I've faced some issues with preAggregations where some measuers were they were being summed instead of being averaged over for `avg` types. Some logic for the `rollup` timeout is a bit confusing.

Although the REST API interface doens't really need any framework, I feel that a simple SDK providing the objects for serialization would be nice to have.

Some more complex charts in the Playground would also be nice to have, together with a simple export like CSV/JSON from the Playground (I do understand that the Playground is... a Playground... but sometimes I use it as an analytics tool to fetch some informations).

I think that we also lack a propert LSP (or a type checker) for Cube on Javascript. Sometimes I makes some mistakes that are related to Cube JS syntax (use of AI on the Code Editor also makes it more prone to error). This would be really helpful. Some features can only be tested once in production (like preAggregations syntax), which makes it a bit harder to integrate. Review collected by and hosted on G2.com.

Using CubeJS Cloud to build our semantic layer has been great! It really just takes a lot of development effort off our shoulders. This is especially great for small teams that don't have DevOps engineers available to help create and monitor a platform. The development and maintenance have been simple to maintain. Their support is also great where I could solve a build issue with a live chat in a short time. There are many features including different endpoints to consume data, which makes embedding analytics into our product straightforward. Review collected by and hosted on G2.com.

I still feel that integration with BI tools is not totally straightforward yet, some query aggregations fail and require some tweaks. I still have to test the new "Semantic Layer Sync" feature. Also, a thing that was very great about dbt metrics that I miss here in CubeJS is the ability to create a new metric based on other metrics as long as they have the same time granularity, CubeJS could find a way to implement something flexible as this as well. Review collected by and hosted on G2.com.

- declarative schema definitions make it easy to built UI interfaces on top of your data

- built-in cache

- extensibility through express middlewares

- great slack community

- easy to get setup Review collected by and hosted on G2.com.

we ran into an issue with cube cloud processes restarting too frequently which meant the entire schema needed to be reloaded on subsequent requests which took nearly 5s and triggered our latency alarms. luckily due to cube's extensibility we were able to implement a custom express middleware to trigger loading of the schema when the instance started up which worked around the problem for us.

the schema compiler is a little quirky and has too much magic which makes it a little more difficult to define schemas in typescript, but not insurmountable.

personally I don't like the online schema editor since we discourage folks from making changes outside of the version control. I wish we could have disabled this feature.

out-of-the-box scripts to scaffold out schemas would be nice. we ended up rolling our own.

cube-sql is interesting, but the lack of ability to do basic math in queries made it unusable for us. Review collected by and hosted on G2.com.

I have been using Cube for the better part of the last 2.5 years. It gave us an out-of-the-box abstraction to quickly slice and dice our data arbitrarily and provide insights and analytics layers for our customers. As Cube has grown, they have added more robust query result caching and pre-aggregation generation. We were fortunate enough that this lined up perfectly with our growth. As we collected more and more data, we knew Cube had our back in ensuring that we could continue to provide performant analytics to our customers. Review collected by and hosted on G2.com.

Setting up more advanced features, such as cube store + pre-aggregations, is very difficult if you have a large dataset. Seeding the aggregations at the scale of millions of data points is no easy feat. You won't find documentation or ready support in that

Cube does provide a Slack channel where you can ask questions from their team and the community; however, it seems to have been deprioritized recently. I will say that the Cube Cloud support team is very responsive. If you use Cube for production workloads, I highly recommend paying for the managed offering after evaluating its fittingness to your use case. Review collected by and hosted on G2.com.

Cube.js has been a game-changer in my development journey, offering an incredible solution for building data-driven applications. With its intuitive data modeling capabilities, I've been able to structure data in a way that truly suits my application's needs. it handles data efficiently, ensuring that my users can explore insights seamlessly. Review collected by and hosted on G2.com.

Cube.js, like any tool with advanced capabilities, has a learning curve. Beginners may find it challenging to grasp all the concepts, especially when dealing with complex data models and SQL queries. However, Cube.js does provide documentation and tutorials to help mitigate this issue. Review collected by and hosted on G2.com.

Easy to make pre-aggregations that are a game-changer in modern data analysis systems, as they are a main layer for best in class optimization. Ease of configuration and deployment was a key for me. Review collected by and hosted on G2.com.

I miss a local/system native optimization program like cloud platform panel - possibility to write preaggs and already check them before deployment, analyze queries and requests to optimize my application even more at data layer. That would be an excellent improvement for general DX (Developer Experience) as well. Review collected by and hosted on G2.com.

Two Cube features that were particularly useful was that Cube is queryable with mutliple APIs : REST, GraphQL and SQL. This saved an enormous amount of time from building out this layer and allowed existing apps to query the Cube layer easily for relevant data. The other one was that the Cube store acts as a caching layer that not only caches queries but also provides very fast aggregations for analytical purposes. This helped create applications which could satisfy a broad range of requirements which made it applicable for both internal and external facing use cases. Review collected by and hosted on G2.com.

For a semantic layer, there are few things that make it feel somewhat feature incomplete. This is by design as cube takes an API first approach preferring to focus on doing few things well. But for users looking for a more "all-in-one" type solution, this is not it. For example, a semantic layer might be expected to have data catalog features or visualization features as well. Cube instead chooses (atleast currently) to faciliate these things rather than providing it out of the box. Review collected by and hosted on G2.com.

Cube makes it incredibly easy to expose our reporting data to our businesses. Our shift to using Cube to serve our reporting data has cut the deployment process of new features and bug fixes from hours to minutes. Also, the onboarding process for getting up to speed with the product was swift, and the Cube team has been very responsive to feature requests. Review collected by and hosted on G2.com.

The only real issue we currently have is with the limit of 50k records that can be served. Some of the datasets that are requested go well above this, and we need to go outside of the tool for a way to service these requests. Review collected by and hosted on G2.com.