Towards a model for InfluxDB Multi-Tenancy

Database instances are now referred to as buckets. The default name we use for a bucket is signals.

Each bucket is attached to an Influx organisation.

An influxdb-writer microservice is associated with a particular InfluxDB organisation.

The influxdb-writer uses a pre-defined API token that is owned by the organisation and allows write-only access to the bucket.

In this way signals from a particular MQTT broker can be written to an influx organisational bucket by a single writer. This creates a mapping between the PIs that are associated with an MQTT broker and an organisation in Influx.

Until such time (possibly never) that a change is made to use InfluxDB dashboards rather than Grafana dashboards then a single InfluxDB user is associated with an InfluxDB organisation.

The exception to this is that the InfluxDB root user is associated with all organisations for maintenance purposes

A Grafana organisation is associated with an InfluxDB organisation through a Grafana DataSource definition. For example the organisations DGOPS_AUTO and DGOPS_CUSTOM would be associated with the dgops organisation in InfluxDB through the credentials supplied to the data source definition. For InfluxDB version 2 these credentials would be an API token that specifically allows read-only access to the named bucket(s) associated with that organisation.

Each Grafana organisation has a default data source that maps through to the InfluxDB bucket for the corresponding InfluxDB organisation

The dashboard-generators are associated with the Grafana organisation through the Grafana API key that they use to access the Grafana API to create dashboards. They specify a default datasource in the dashboard which is defined for the organisation.

Grafana users are attached to organisations which allows access to dashboards and their associated data