Status

March 24, 2026 · View on GitHub

Last updated: March 2026

Overview

ComponentLocationStatusNotes
Core Specificationspec/v1/specification.md✅ StableIncludes ai-gate type
JSON Schemaspec/v1/schema.json✅ StableValidation schema
Judgment SLOsspec/v1/judgment-slos.md✅ DocumentedDetailed reference
GitHub Actionaction/✅ AvailableCI/CD validation
NthLayerExternal repo🟡 AlphaReliability-as-code CLI
nthlayer-learn (Verdict)External repo✅ ImplementedData primitive library
nthlayer-measureExternal repo✅ ImplementedQuality measurement + governance
nthlayer-correlateExternal repo✅ ImplementedPre-correlation agent
nthlayer-respondExternal repo✅ ImplementedMulti-agent incident response
Change Eventsconventions/change-events/📝 DraftedOTel proposal
Decision Telemetryconventions/decision-telemetry/📝 DraftedOTel proposal

Legend:

  • ✅ Stable / Implemented
  • 🟡 Alpha / In Progress
  • 📝 Drafted / Needs Review

Specification (spec/)

Core Specification

Status: ✅ Stable

The foundation. Defines ServiceReliabilityManifests including:

  • Service metadata (name, tier, labels)
  • Service types (api, worker, stream, ai-gate)
  • SLO definitions (availability, latency, throughput)
  • Dependencies (upstream/downstream, expectations)
  • Ownership (team, contacts)
  • Contracts (external promises)

Location: spec/v1/specification.md

ai-gate Extension

Status: ✅ Stable

Extension for AI decision services. Included in core spec.

  • type: ai-gate declaration
  • Judgment SLO fields
  • Instrumentation requirements

Location: spec/v1/specification.md (Section 5)

Judgment SLOs Reference

Status: ✅ Documented

Detailed reference for judgment SLO metrics:

  • Reversal rate
  • High-confidence failure rate
  • Calibration (ECE)
  • Audit layers (proactive spot-checks)
  • Maturity levels

Location: spec/v1/judgment-slos.md

JSON Schema

Status: ✅ Stable

Machine-readable validation schema.

Location: spec/v1/schema.json

Conventions (conventions/)

Change Events

Status: 📝 Drafted

OTel semantic conventions for operational changes.

ArtifactStatus
SpecificationDrafted
Implementation planDrafted
ExamplesDrafted

Next steps:

  1. Engage OTel CI/CD SIG
  2. Socialize in #cicd-o11y Slack
  3. Submit PR to semantic-conventions repo

Location: conventions/change-events/

Decision Telemetry

Status: 📝 Drafted

OTel semantic conventions for AI/human decisions.

ArtifactStatus
gen_ai.decision.* attributesDrafted
gen_ai.reversal.* attributesDrafted
gen_ai.outcome.* attributesDrafted
ExamplesDrafted

Next steps:

  1. Continue OTel GenAI SIG engagement
  2. Align with genai-utils work
  3. Submit PR

Location: conventions/decision-telemetry/

Components (components/)

NthLayer

Status: 🟡 Partial

Reference implementation of OpenSRM. Generates monitoring artifacts from manifests.

CapabilityStatus
Manifest validation✅ Done
Prometheus rule generation✅ Done
Grafana dashboard generation✅ Done
PagerDuty config generation✅ Done
Topology export🟡 Partial
Judgment SLO alerting❌ Not started

Location: External repo (github.com/rsionnach/nthlayer)

Next:

  • Export topology for nthlayer-correlate consumption
  • Add judgment SLO alert generation

nthlayer-correlate

Status: ✅ Implemented (Phase 2 Tier 1)

Pre-correlation agent for AI-native observability.

ArtifactStatus
ProposalComplete
Technical appendixComplete
Snapshot schemaComplete
Correlation engineImplemented
Ingestion layerImplemented
State machineImplemented
SQLite event storeImplemented
CLI (serve/status/replay)Implemented
Tests (14 files, ~111 tests)Passing
Synthetic scenarios (5)Complete

Location: External repo (github.com/rsionnach/nthlayer-correlate)

Package: pip install nthlayer-correlate (0.1.0a1)

nthlayer-respond

Status: ✅ Implemented (Phase 3 complete)

Multi-agent incident response system.

ArtifactStatus
ArchitectureComplete
Coordinator (state machine)Implemented
Triage agentImplemented
Investigation agentImplemented
Communication agentImplemented
Remediation agentImplemented
Safe actions registryImplemented
Context store (SQLite)Implemented
CLI (serve/status/replay/approve/reject)Implemented
Tests (13 files, ~168 tests)Passing
Synthetic scenarios (8)Complete

Location: External repo (github.com/rsionnach/nthlayer-respond)

Package: pip install nthlayer-respond (0.1.0a1)

Tooling (action/)

GitHub Action

Status: ✅ Available

Validates OpenSRM manifests in CI pipelines.

- uses: rsionnach/opensrm@v1
  with:
    manifest: service.reliability.yaml

Location: action/, action.yml

Current Focus

Completed (Q1 2026):

  • nthlayer-learn (Verdict) — Python library + CLI
  • nthlayer-measure — Phase 1 verdict integration, governance
  • nthlayer-correlate — Phase 2 Tier 1 implementation
  • nthlayer-respond — Phase 3 full implementation
  • nthlayer.io website with live topology demo

Next:

  • OTel decision telemetry PR
  • Change events OTel PR
  • Production deployment hardening

Decisions Log

DateDecisionRationale
2026-02-20Expand opensrm to include ecosystemSingle repo for spec + vision
2026-02-20Rename "Situations" to "nthlayer-correlate"Avoid Moogsoft/BigPanda terminology
2026-02-20Hybrid generation model for nthlayer-correlateBatch + incident-triggered + refresh
2026-02-16Add audit layer to judgment SLOsReversal rate alone misses unreversed bad decisions
2026-01-23Use OpenSRM nameFollows OpenSLO, OpenTelemetry pattern
2026-01-22Position change events as OTel extensionComplement CDEvents, not compete

Roadmap

Phase 1: Specifications (Q1 2026) — Complete

  • Core OpenSRM spec
  • ai-gate extension
  • Judgment SLOs reference
  • Change events OTel proposal (drafted)
  • Decision telemetry OTel proposal (drafted)

Phase 2: Core Implementations (Q1 2026) — Complete

  • nthlayer-learn (Verdict) — Python library + CLI
  • nthlayer-measure — evaluation pipeline + verdict integration + governance
  • nthlayer-correlate — correlation engine, ingestion, snapshots, state machine
  • nthlayer-respond — coordinator, 4 agents, safe actions, crash recovery

Phase 3: Integration & Hardening (Q2 2026)

  • OTel convention PRs (CI/CD SIG, GenAI SIG)
  • NthLayer judgment SLO support
  • End-to-end integration testing across components
  • Production deployment guides

Phase 4: Adoption (Q3-Q4 2026)

  • Production deployments
  • Community contributions
  • Conference talks