We've introduced (in beta) Tinybird Self-Managed: A deployable version of Tinybird that runs in your own cloud environment, giving you complete control over your analytical backend while keeping everything developers love about Tinybird—speed, scalability, and ease of use.
With Tinybird Self-Managed, you can deploy Tinybird inside your AWS infrastructure (with GCP and Azure support coming soon), integrate it with your existing private data sources, and manage performance, scaling, and upgrades on your terms.
Why self-managed?
Tinybird started as a cloud-first platform. We wanted to eliminate operational overhead for developers and allow them to build and scale real-time data applications without managing infrastructure. This worked well for many teams, but we heard consistent feedback from customers who needed more control over their data environments.
While Tinybird Cloud remains the best option for most users, we recognize that some teams require self-hosted deployments for valid reasons:
- Data sovereignty & compliance: You need to keep your data within your security perimeter and specific geographic regions to meet regulatory requirements—e.g., GDPR or HIPAA.
- Deeper integration: You want to connect Tinybird directly to your existing private data sources and internal systems without moving data across networks.
- Infrastructure optimization: You want to tune hardware resources for your specific workload patterns and tailor performance based on your specific business needs.
- Cost efficiency: You want to reduce egress costs by keeping data processing within your network, and you want to pay for storage costs with your existing cloud commitments.
- Operational control: You want to manage performance, scaling, and upgrades on your terms.
Instead of installing ClickHouse® open source or a different database solution, Tinybird Self-Managed provides you with the OLAP database, ingestion APIs, API gateway, observability, and backpressure mechanisms needed to prevent hardware overload while giving you more fine-tuned control over infrastructure and resources.
How does it work?
Tinybird Self-Managed runs as a containerized deployment in your cloud environment, using Kubernetes to provide a simple yet scalable setup. You can deploy it with a single command:
This generates a set of files to deploy your Tinybird Self-Managed to AWS (soon GCP and Azure) using a Kubernetes cluster, and it sets up everything needed to run Tinybird in production.
Once deployed, you can log in to cloud.tinybird.co and you'll be able to select the region and work on it. (btw, you can also connect the UI to your local image).
Behind the scenes, it is the very same Tinybird Local container with support for data persistence. While this is currently not enough for medium and large deployments, as it lacks HA and other enterprise features, it covers 50% of self-managed use cases perfectly.
What’s next?
We’re starting with a single-node deployment (ideal for testing and smaller workloads), but our roadmap includes:
- Expanded cloud support: GCP and Azure support in addition to AWS.
- Efficient cloud storage: ClickHouse open source struggles with S3 persistence, but our Tinybird ClickHouse fork optimizes Cloud Storage, reducing unnecessary S3 operations.
- Battle-tested HA and scalability: The same high concurrency, high throughput data ingestion capabilities and low latency real-time query performance as Tinybird Cloud.
- Automated upgrades: Zero downtime, simple, and easy upgrades.
- Backups: Cost-efficient incremental backups.
- Advanced monitoring & alerting: Prometheus exporters and pre-configured Grafana dashboards for real-time monitoring.
Try it today
While in beta, Tinybird Self-Managed is free to use, and small deployments will likely remain free forever.
If you need full control over your real-time data platform, Tinybird Self-Managed is now available to try. Get started today:
Or check out the Tinybird Self-Managed regions documentation for more details.
FAQ
Technical Requirements
Before getting started with Tinybird Self-Managed, ensure you have a Kubernetes cluster in AWS that meets the following minimum hardware requirements
- CPU: 4 cores (8+ recommended for production)
- RAM: 16GB (32GB+ recommended for production)
- Storage: 100GB SSD storage (scalable based on data volume)
How to scale?
Currently, you can add more capacity by vertically scaling your node to increase its size. This approach still requires some configuration fine-tuning, which we are working on as well.
Cost estimation
While Tinybird Self-Managed is free during beta, you'll of course need to consider the infrastructure costs (EC2/VM instances, EBS/Storage, Network, K8s cluster). A typical monthly cost for a standard setup:
- Small (4 cores, 16GB RAM): $150-300/month
- Medium (8 cores, 32GB RAM): $300-600/month
- Large (16+ cores, 64GB+ RAM): $600+/month
