Document Alloy to Cortex Tenant to Mimir integration

General MR

Summary

• Keep monitoring's Prometheus-disabled behavior: disable token-dependent control-plane ServiceMonitors by default while retaining kube-state-metrics and node-exporter for Alloy scraping.
• Keep the built-in Alloy/monitoring metrics path working without Prometheus, including exporter scraping and matching policies.
• Remove the umbrella's Cortex Tenant-specific API, remote-write destination, network policies, and Flux dependency logic. Cortex Tenant remains a custom package.
• Provide a tested Alloy → Cortex Tenant → Mimir example values file with explicit destination, tenant label, network policies, and dependencies.
• Pin Alloy 4.3.2-bb.2 and Mimir 6.2.1-bb.1 for both Git and Helm sources. The example pins custom Cortex Tenant to 0.8.1-bb.2.

Relevant logs/screenshots

• helm lint ./chart, package schema and values reference checks, focused Helm unit tests (27 tests), and values-overlay Bats tests passed.
• The example values and the updated Cortex Tenant package CI overlay rendered against this branch; Alloy selects Cortex Tenant and Mimir distributor ingress is enabled explicitly.
• The full local Helm unit run had 509 passes and 13 unrelated post-render errors caused by the protected template cache or a package fetch failure. External clean-install CI remains the cross-package evidence.

Linked Issue

Closes #3464 (closed)

Package MRs

• Alloy !190
• Mimir !128
• Cortex Tenant !2
• Cortex Tenant CI overlay !4

Upgrade Notices

Cortex Tenant is now an available Big Bang package. Cortex Tenant is not included in the umbrella but follow the example values file linked to setup a working integration with Alloy and Mimir.

Cortex Tenant is not installed or configured automatically by Big Bang. To use Alloy → Cortex Tenant → Mimir, apply the example values, choose the tenant label, and adjust the transport for ambient or non-Istio deployments. For a monitoring installation without Prometheus, set monitoring.prometheusMetrics.enabled: false and monitoring.values.upstream.prometheus.enabled: false (or the corresponding packages.monitoring values with packageConfiguration.version: v1).