Tracking issue: Mimir distributed deployment

[lucabello]

Deployment Graph

Workloads:

• coordinator: nginx + Grafana agent
• worker: Mimir

graph LR

subgraph mimir
worker1 ---|mimir-cluster| coordinator+gagent
worker2 ---|mimir-cluster| coordinator+gagent
worker3 ---|mimir-cluster| coordinator+gagent


end

subgraph cos-lite
coordinator+gagent ---|logging| loki
coordinator+gagent ---|remote-write| prometheus
coordinator+gagent ---|dashboards| grafana
coordinator+gagent ---|trace| tempo

end

Loading

Alert rules and dashboards belong to the coordinator.

Issues to complete

Mimir Worker

• rework the charm to the new design mimir-worker-k8s-operator#11
• add TLS support mimir-worker-k8s-operator#12

Mimir Coordinator

• rework the charm to the new design mimir-coordinator-k8s-operator#13
• Add the nginx workload mimir-coordinator-k8s-operator#14
• add TLS support mimir-coordinator-k8s-operator#15
• Add self-monitoring integrations mimir-coordinator-k8s-operator#16

Others

• Create a relation interface for Mimir distributed charm-relation-interfaces#120

[sed-i]

Potential experimentation with the mimir charms:

• Roll our own "event groups" (using stored state)
• Manifest driven

[lucabello]

About the relation databags, the worker should provide:

• its roles (read from the config options);
• its internal address: we should use the highest priority address as we do with all of our charms (FQDN first, bind address second);
• the port which the workload listens to;
• the scheme (http/https);
• its topology (model and unit name?).

The coordinator should provide:

• the Mimir configuration.

---

This list of fields should be expanded/changed as we find new issues or features to address.

[lucabello]

This tracking issue isn't useful anymore, closing :)