3. Deploy an example instance#
For production grade deployments we recommend to use Kubernetes. You can find a set of Helm Charts we use for that here. Official charts of services such as PostgresSQL and MongoDB can be found in their repos). You would need around 6 kubernetes worker nodes with 8GB ram and 2 CPUs for a small HA deployment (excluding analytics services). Optimal sizing depends on the 1) volume and 2) frequency of data you need to inject in the platform; 3) number of side services (e.g. monitoring) and 4) desired availability.
Deploying production ready services require quite some time and expertise in using K8S, that are beyond what can be offered by our team to the community. If you are interested, Martel offers consulting services to help in the process.