Self-hosting Novu involves setting up and maintaining your communication infrastructure using your own servers. While this offers full control and customization, some features which are available only for Novu’s cloud-managed won’t be available in a self-hosted environment.

If you require guidance on setting up a production-scale configuration, official IaaS recommendations, please don’t hesitate to contact our team. They can provide assistance regarding enterprise edition licenses or cloud-prem deployment to meet your specific needs.

​
Minimum System Requirements

For basic installation, a server with 16GB RAM and 8 CPUs is fine. However, specifications vary depending on requirements and here are sample specifications for handling 1M plus notifications:

  • If everything is running on one machine, then at least 36 vCPU and 64GB of RAM.
  • If services are running across containers, then it is recommended to run 3 containers per service, each container having 2 vCPU and 4GB of RAM.
  • If you’re using Kubernetes, then the K8 system needs to have 36 vCPU and 64GB of RAM at least in the cluster to run the servers.
  • On AWS, a c6g.8xlarge instance with 32 vCPU and 64GB of RAM.
  • In all cases, you need to also have 2 Redis clusters with the queue one having an Append only log file active (AOL) to prevent instances of job loss during an outage, a MongoDB cluster equivalent to a M20 or greater on Atlas, and 10GB of S3 storage.
  • Lastly, the queue Redis instance should be 8GB MINIMUM as there needs to be enough space to handle spikes in notifications or the whole system will brick.
Note that these specifications are subject to your requirements, the scale of your operations, and should be adjusted as needed.
Github login is not available in self-hosting. If you’re self hosting Novu, use your email and password to log into your Novu account!

Was this page helpful?

WhatsAppDeploy with Docker