Deploy Mimir using the provided Helm chart (grafana/mimir-distributed) and configure an object storage backend (S3, GCS, or Azure Blob) in the mimir.yaml values file
Set the compactor, store-gateway, and ingester component replicas appropriate to your scale; start with the small.yaml reference values file from the Mimir repository
Configure Prometheus remote_write to send metrics to the Mimir distributor endpoint at http://mimir-distributor:8080/api/v1/push, adding an X-Scope-OrgID header for tenant isolation
Add a Grafana data source of type Prometheus pointing to the Mimir query-frontend endpoint and set the HTTP header X-Scope-OrgID to the tenant ID
Verify multi-tenant isolation by querying the same metric from two different tenants and confirming data does not bleed across tenant boundaries
Configure ruler storage in Mimir so that existing Prometheus alerting rules and recording rules can be managed via the Mimir ruler API (/api/v1/rules)
Known gotchas
Mimir's ingesters use WAL (write-ahead log) for durability; ensure the PersistentVolumeClaim size is sufficient for your ingestion rate or ingesters will crash-loop during WAL replay
The X-Scope-OrgID header must be set consistently in both remote_write and Grafana data source; a mismatch means queries return empty results with no error
Object storage bucket permissions must allow both read and write for all Mimir components; the compactor and store-gateway require separate IAM paths in some cloud configurations
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp