Skip to main content


Client binaries

When installing Otomi using the chart, make sure the following client binaries exist:

  • Kubectl to access the cluster
  • Helm for helm chart installation of Otomi


AWS, Azure and Google

Otomi supports 3 cloud provides. The aws, azure and google providers include creating optimized storage classes and optional integration with L7 load balancers (like an ALB in AWS and a Application Gateway in Azure).

Digital Ocean

Install Otomi on a Kubernetes cluster in Digital Ocean, use the digitalocean provider.

All others

Use the custom provider for all other clouds and when running Kubernetes on your own hardware (including Minikube on your local machine). The custom provider uses the default available storage classes. The only requirement for using the custom provider is to be able to create a Kubernetes LoadBalancer Service that obtains an external accessible IP.

Kubernetes versions

Otomi currently supports the following Kubernetes versions:

  • 1.18
  • 1.19
  • 1.20
  • 1.21
  • 1.22
  • 1.23

Minimal compute resource requirements

Otomi requires a node pool with at least 6 vCPU threads and 8GiB+ RAM. Note that this is the requirements for a minimal (default) install. When activating more apps, you'll probably need more resources.


The minimal resource requirement to run Otomi is based on running Core Apps only! The core apps provide an advanced ingress architecture based on Nginx, Istio, Keycloak, Oaut2 Proxy and Certmanager. Activating optional apps will require more compute resources. We advise to have a node pool available with 12 vCPU and 32 GiB memory.


To use the network policies feature in Otomi, make sure to install the Calico CNI or any other CNI that supports Kubernetes network polices.