Value capture - outcome value and unit economics with flipped denominators

Description

The second half of the origin question: capture the value the platform brings, so the conversation moves to upside rather than cost.

Outcome value - what observability did

From grafanacloud-usage:

• pages raised per observed service and per team;
• MTTA and MTTR at p50 only, with observations above the 3600-second finite bucket expressed as a count;
• ownership completeness through the No team and No service shares;
• every alert-group ratio restricted to the stacks that report response timing;
• a missing histogram observation described precisely as no acknowledgement recorded.

From Pillar J:

• dashboard opens and distinct viewers, retaining its explicit window and per-stack de-duplication caveat.

Unit economics, denominators flipped

Use Grafana-published billing spend directly instead of reconstructing charges from quantities and contract rates. Show clearly-labelled product-specific and estate-wide ratios including spend per observed service, Application Observability service, Infrastructure host/container hour, acknowledgement, billed user and dashboard viewer. Custom rate cards remain only for currency that must be derived from a non-currency quantity. Unknown or partial pricing is absent, never zero or presented as an estate total.

Service observability completeness replaces the earlier protected-service wording. Metrics, logs, traces, profiles, explicit service dashboard, alert and SLO remain visible as seven components; deployment-configurable weights affect only the S3 score. Active direct alert routing remains a separate flag.

What this is not

Not a link to ARR or pipeline. That join belongs outside this repository. This surface exposes the existing org and stack join keys and stops.

Acceptance Criteria

• #1 MTTA/MTTR present p50 only, with the tail expressed as a count above the top finite bucket
• #2 Every OnCall ratio restricts its denominator to the timing-reporting population
• #3 Ownership completeness rendered from the No team and No service shares
• #4 Unit-economics panels use billed users for money, never active users
• #5 Grafana-published billing spend may be rendered directly; custom rate cards are required only when deriving currency from non-currency quantities, and unknown or partial pricing is never presented as zero or as an estate total
• #6 No ARR, pipeline or account data is joined in this repository

Definition of Done

• #1 python3 -m pytest tests -q
• #2 tofu fmt -check -recursive terraform; tofu init -backend=false and tofu validate pass for terraform/ and terraform/examples/standalone/
• #3 customer-identifier and shipped-text gates from .github/workflows/ci.yml return clean

Implementation Plan

1. Extend the atomic Mimir signal inventory with an explicitly windowed, bounded SLO service matcher and prove success, empty, truncation and failure semantics first.
2. Add deployment-configurable seven-component observability completeness weights with validation; calculate only fully known per-service scores in the S3 register and keep routed-alert state separate.
3. Add Coverage outcome-value and unit-economics panels using Grafana-published overage metrics, timing-population-safe OnCall queries, Pillar J viewer data and billed-user denominators.
4. Re-derive hydration dependencies and generated artifacts, validate dashboards and Terraform, run the full repository gates and CodeRabbit review.
5. Create a follow-up Backlog task for live demo-estate validation, then finalize, commit and push GCI-0008.07.

Implementation Notes

Implemented atomic Mimir SLO service matching, configurable seven-component completeness scoring in the S3 register, population-safe OnCall outcome panels, and Grafana-published billing unit economics including Application Observability. CodeRabbit completed with one minor test-strengthening suggestion, which was applied. Verification: 1,388 tests passed, 2 skipped, 6,614 subtests; OpenTofu formatting and fresh-cache validation passed for the module and standalone example; all dashboards assembled offline. Live snapshot remains for GCI-0008.08 against an explicitly selected development deployment. The private identifier pattern is CI-only, so DoD #3 remains open until pushed CI passes.

Final Summary

Completed Pillar K value capture: atomic SLO association, configurable seven-component service observability completeness, timing-population-safe OnCall outcomes, and Grafana-published billing unit economics including Application Observability. Verified by 1,388 tests plus 6,614 subtests, offline assembly of all eleven dashboards, fresh-cache Terraform validation, CodeRabbit and the secret-backed CI identifier gate. Live development validation is GCI-0008.08.