SCOPE

This file helps you quickly understand what this repository is about, when it is relevant, and when it is not. It is intentionally lightweight and may be incomplete.

One-liner

S3 Platform Services layer of the Railiance OAS Stack — owns shared cluster services: PostgreSQL HA, Valkey cache, secret management, identity integration, and object storage.

Core Idea

Railiance is structured as five independent repos per OAS Stack layer. This repo is S3 — the platform services that multiple applications share. PostgreSQL HA (repmgr + pgpool) and Valkey (Redis-compatible cache) are the first services being extracted here, moved from the Gitea subchart in S2 into standalone Helm releases under S3 so the boundary rule is properly enforced.

In Scope

PostgreSQL HA (repmgr + pgpool) as a standalone Helm release
Valkey / Redis-compatible cache as a standalone Helm release
Secret management infrastructure
Identity services integration point (with net-kingdom)
Message brokers (RabbitMQ, similar)
Object storage (MinIO / S3-compatible)
Backup and recovery services for platform data

Out of Scope (per ADR-003)

OS-level concerns → railiance-infra (S1)
Kubernetes runtime → railiance-cluster (S2)
Developer tooling, CI/CD → railiance-enablement (S4)
Application deployments → railiance-apps (S5)
No re-configuration of S1/S2 concerns from this repo

Relevant When

Deploying or managing shared services that multiple S5 applications depend on
Extracting platform services from application Helm subcharts (boundary enforcement)
S2 cluster is operational and platform layer can now be established

Not Relevant When

S2 (cluster runtime) is not yet operational (pre-condition not met)
Application-specific database schemas or migrations (those belong in S5 apps)
Infrastructure or cluster work (wrong layer)

Current State

Status: active / emerging
Implementation: PostgreSQL HA + Valkey extraction from S2 in progress (RAIL-PL-WP-0001)
Stability: emerging — first platform workplan active
Usage: shared database and cache services for all S5 applications

How It Fits

Upstream dependencies: railiance-cluster (S2) — k3s running, Helm available, smoke tests passing
Downstream consumers: railiance-enablement (S4), railiance-apps (S5) — all depend on platform services
Often used with: net-kingdom (identity services integration), railiance-cluster (prior layer)

Terminology

Preferred terms: OAS Stack Level S3, platform services, boundary rule, migration (extracting from S2 subcharts)
Potentially confusing terms: "migration" here means moving Helm releases between layers, not database schema migration

railiance-cluster (S2) — pre-condition; PostgreSQL was previously managed here (being extracted to S3)
railiance-apps (S5) — consumes database and cache services from S3
net-kingdom — identity services integration point at the platform layer

Provided Capabilities

type: infrastructure
title: PostgreSQL HA (repmgr + pgpool)
description: Highly available PostgreSQL cluster using repmgr for replication and pgpool for connection pooling — shared database service for all platform applications.
keywords: [postgresql, postgres, ha, repmgr, pgpool, database, replication]

type: infrastructure
title: Valkey / Redis-compatible cache
description: Shared Redis-compatible cache service (Valkey) for all applications in the Railiance stack.
keywords: [valkey, redis, cache, shared, session, queue]

type: data
title: Object storage (MinIO / S3-compatible)
description: S3-compatible object storage service (MinIO) for artifact storage, backups, and large file handling across platform applications.
keywords: [minio, s3, object-storage, storage, artifacts, backup]

Getting Oriented

Start with: CLAUDE.md (session protocol, boundary rules)
Key files / directories: workplans/RAIL-PL-WP-0001-platform-baseline.md, helm/ (platform Helm charts), Makefile
Pre-conditions: railiance-cluster (S2) converged with k3s running; active backup on Nextcloud before migration steps

4.2 KiB Raw Blame History