IoTDB Cloud

This is a proof of concept for a self service hosted IoTDB Cloud Solution. The basic Idea is to get a Service like MongoDBs Atlas but for timeseries based on IoTDB.

Functionality

The general idea is to have a setup where new Apache IoTDB instances can be created by users or interested parties. Current Implementation allows to create an IoTDB Instance with a single click. In the background in Kubernetes the following ressources are created:

• IoTDB Single Node Instance (as stateful set with PVC for data)
• Corresponding Headless Service
• Service which exposes IoTDB as External IP
• Initializer Job to configure the IoTDB instance (currently only set password)

Look and Feel

Roadmap

• First prototyp
• User Login
• Rich Settings for Instance creation
• Move Async Tasks to Celery
• Limit Demo Accounts for Users
• Send Email when instance is ready
• Use TCP with Ingress to reduce external IPs that are needed
• Use Proper Operator for Resource management (via CRD) instead of template files

Installation in a Kubernetes Cluster

• First, you need a running database (e.g. postgres)
• And a provisioner for the certificates
• Modify the namespace in rbac.yaml to create a suitable service user (change namespace)
• Run it via kubectl apply -n <your-namespace> -f rbac.yaml
• Modify the kubernetes.yaml accordingly (the ConfigMap) and service user from above, if necessary
• Run it via kubectl apply -n <your-namespace> -f kubernetes.yaml
• If wanted, adopt and apply ingress.yaml to configure an ingress ressource for the portal service

Have fun!