The simplest way
to run Spark

Just send your code, we handle the mechanics.

Want to learn more? Data Mechanics is available on any cloud provider.  Reach out so we give you a live demo and prepare a deployment for you.

Get a Demo
Maintenance is on us

Our platform tunes the infrastructure parameters and Spark configurations dynamically and continuously for each of your workloads, making your applications 2x faster and more stable on average. We make using Spark as easy as it should be.

The fastest experience with Spark

Notebooks and scheduled jobs are triggered and autoscale in seconds to react to the desired load. No more waiting 10 minutes for a cluster to spin up!

Open-Source and Kubernetes native

Bring the DevOps best practices to your data stack. Kubernetes-native, our platform is available on any cloud provider and integrates with the tools you already use and love. Just point your notebooks solution to our gateway and you’re ready to go.

Frequently Asked Questions

How do I use it?

Run autoscaled Jupyter kernel with Spark support from the notebook environment you already have. Use our operator library to launch scheduled jobs from your favorite orchestrator (Airflow, Luigi, Azkaban, custom schedulers).

We automatically set the infrastructure parameters and Spark configurations so you don’t need to worry about them.

What’s the installation process like?

Provide us with permissions scoped to the desired cloud account and region and our service will deploy the platform in minutes. We manage the Kubernetes cluster and deployment for you, all you need to do is point your notebooks and jobs to our gateway and you're ready to go. 

If you want the flexibility to install the platform yourself, it can also be installed on an existing Kubernetes cluster through a Helm chart.

How much does it cost?

Data Mechanics charges a pay-as-you-go fee based on the total computation time spent in Spark applications, billed monthly, without commitment. This fee is additional to your cloud provider costs.

All the competing data platforms base their fee on the total server uptime, whether these servers are actually used by Spark or whether they are sitting idle. As a result, these platforms have no incentive to reduce the wasted idle time, and up to 80% of your monthly bill typically comes from these wastes.

At Data Mechanics, we don’t charge you on this idle time so that we’re incentivized to make your data infrastructure as performant and cost-effective as possible.

Which infrastructures do you support?

Our platform is available on Google Cloud Platform GKE, Amazon Web Services EKS, Azure AKS, and your own Kubernetes setup.

Do you support other big data technologies than Spark?

We currently focus on Spark, but our vision is to expand beyond Spark and make the best data science and data processing technologies available on our containerized data platform. Let us know which tech you're interested in.

The simplest way to run Apache Spark
Contact us for a demo or to learn more about Data Mechanics!
Get a Demo