Add tests for SBOM generation determinism across multiple formats

- Created `StellaOps.TestKit.Tests` project for unit tests related to determinism.
- Implemented `DeterminismManifestTests` to validate deterministic output for canonical bytes and strings, file read/write operations, and error handling for invalid schema versions.
- Added `SbomDeterminismTests` to ensure identical inputs produce consistent SBOMs across SPDX 3.0.1 and CycloneDX 1.6/1.7 formats, including parallel execution tests.
- Updated project references in `StellaOps.Integration.Determinism` to include the new determinism testing library.
This commit is contained in:
master
2025-12-23 18:56:12 +02:00
committed by StellaOps Bot
parent 7ac70ece71
commit 491e883653
409 changed files with 23797 additions and 17779 deletions

View File

@@ -1,133 +1,62 @@
# StellaOps — HighLevel Architecture
# Architecture Overview (High-Level)
<!--
Use constants injected at build:
{{ dotnet }} = "10 LTS"
{{ angular }} = "20"
-->
This document is a high-level orientation to StellaOps: what components exist, how they fit together, and what "offline-first + deterministic + evidence-linked decisions" means in practice.
This document offers a birdseye view of how the major components interact,
why the system leans *monolithplusplugins*, and where extension points live.
For the full reference map (services, boundaries, detailed flows), see `docs/07_HIGH_LEVEL_ARCHITECTURE.md`.
> For a *timeline* of when features arrive, see the public
> [roadmap](/roadmap/) — no version details are repeated here.
## Guiding Principles
---
- **SBOM-first:** scan and reason over SBOMs; fall back to unpacking only when needed.
- **Deterministic replay:** the same inputs yield the same outputs (stable ordering, canonical hashing, UTC timestamps).
- **Evidence-linked decisions:** policy decisions link back to specific evidence artifacts (SBOM slices, advisory/VEX observations, reachability proofs, attestations).
- **Aggregation-not-merge:** upstream advisories and VEX are stored and exposed with provenance; conflicts are visible, not silently collapsed.
- **Offline-first:** the same workflow runs connected or air-gapped via Offline Kit snapshots and signed bundles.
## 0·Guiding principles
## System Map (What Runs)
| Principle | Rationale |
|-----------|-----------|
| **SBOMfirst** | Scan existing CycloneDX/SPDX if present; fall back to layer unpack. |
| **Δprocessing** | Reanalyse only changed layers; reduces P95 warm path to\<5s. |
| **Allmanaged code** | Entire stack is 100% managed (.NET / TypeScript); no `unsafe` blocks or native extensions — eases review and reproducible builds. |
| **Restarttime plugins** | Avoids the attack surface of runtime DLL injection; still allows custom scanners & exporters. |
| **Sovereignbydesign** | No mandatory outbound traffic; Offline Kit distributes feeds. |
At a high level, StellaOps is a set of services grouped by responsibility:
---
- **Identity and authorization:** Authority (OIDC/OAuth2, scopes/tenancy)
- **Scanning and SBOM:** Scanner WebService + Worker (facts generation)
- **Advisories:** Concelier (ingest/normalize/export vulnerability sources)
- **VEX:** Excititor + VEX Lens (VEX observations/linksets and exploration)
- **Decisioning:** Policy Engine surfaces (lattice-style explainable policy)
- **Signing and transparency:** Signer + Attestor (DSSE/in-toto and optional transparency)
- **Orchestration and delivery:** Scheduler, Notify, Export Center
- **Console:** Web UI for operators and auditors
## 1·Module graph
## Infrastructure (What Is Required)
```mermaid
graph TD
A(API Gateway)
B1(Scanner Core<br/>.NET latest LTS)
B2(Concelier service\n(vuln ingest/merge/export))
B3(Policy Engine OPA)
C1(Redis 7)
C2(PostgreSQL 16)
D(UI SPA<br/>Angular latest version)
A -->|gRPC| B1
B1 -->|async| B2
B1 -->|OPA| B3
B1 --> C1
B1 --> C2
A -->|REST/WS| D
```
**Required**
---
- **PostgreSQL:** canonical persistent store for module schemas.
- **Valkey:** Redis-compatible cache/streams and DPoP nonce store.
- **RustFS (or equivalent S3-compatible store):** object storage for artifacts, bundles, and evidence.
## 2·Key components
**Optional (deployment-dependent)**
| Component | Language / tech | Responsibility |
| ---------------------------- | --------------------- | ---------------------------------------------------- |
| **API Gateway** | ASP.NET Minimal API | Auth (JWT), quotas, request routing |
| **Scanner Core** | C# 12, Polly | Layer diffing, SBOM generation, vuln correlation |
| **Concelier (vulnerability ingest/merge/export service)** | C# source-gen workers | Consolidate NVD + regional CVE feeds into the canonical PostgreSQL store and drive JSON / Trivy DB exports |
| **Policy Engine** | OPA (Rego) | admission decisions, custom org rules |
| **Redis 7** | KeyDB compatible | LRU cache, quota counters |
| **PostgreSQL 16** | JSONB storage | SBOM & findings storage |
| **Angular {{ angular }} UI** | RxJS, Tailwind | Dashboard, reports, admin UX |
- **NATS JetStream:** optional messaging transport in some deployments.
- **Transparency log services:** Rekor mirror (and CA services) when transparency is enabled.
---
## End-to-End Flow (Typical)
## 3·Plugin system
1. **Ingest evidence sources:** Concelier and Excititor ingest upstream advisories/VEX into immutable observations with provenance.
2. **Scan:** Scanner accepts an SBOM or image reference, produces scan facts and evidence artifacts.
3. **Decide:** Policy evaluation merges scan facts with advisory/VEX evidence to produce an explainable verdict.
4. **Seal:** Signer/Attestor wrap outputs into signed bundles (DSSE/in-toto) and optionally anchor in transparency logs.
5. **Export and notify:** Export Center produces portable evidence bundles and Offline Kit material; Notify delivers digests/incidents.
6. **Operate:** Console exposes triage, explainability, verification, and governance workflows.
* Discovered once at startup from `/opt/stella/plugins/**`.
* Runs under Linux user `stellaplugin` (UID1001).
* Extension points:
## Extension Points (Where You Customize)
* `ISbomMutator`
* `IVulnerabilityProvider`
* `IResultSink`
* Policy files (`*.rego`)
* Each DLL is SHA256 hashed; digest embedded in the run report for provenance.
- **Scanner analyzers** (restart-time plug-ins) for ecosystem-specific parsing and facts extraction.
- **Concelier connectors** for new advisory sources (preserving aggregation-only guardrails).
- **Policy packs** for organization-specific gating and waivers/justifications.
- **Export profiles** for output formats and offline bundle shapes.
Hotplugging is deferred until after v1.0 for security review.
## References
---
## 4·Data & control flow
1. **Client** calls `/api/scan` with image reference.
2. **Gateway** enforces quota, forwards to **Scanner Core** via gRPC.
3. **Core**:
* Queries Redis for cached SBOM.
* If miss → pulls layers, generates SBOM.
* Executes plugins (mutators, additional scanners).
4. **Policy Engine** evaluates `scanResult` document.
5. **Findings** stored in PostgreSQL; WebSocket event notifies UI.
6. **ResultSink plugins** export to Slack, Splunk, JSON file, etc.
---
## 5·Security hardening
| Surface | Mitigation |
| ----------------- | ------------------------------------------------------------ |
| Container runtime | Distroless base, nonroot UID, seccomp + AppArmor |
| Plugin sandbox | Separate UID, SELinux profile, cgroup 1 CPU /256MiB |
| Supply chain | Cosign signatures, intoto SLSA Level3 (target) |
| Secrets | `Docker secrets` or K8s `Secret` mounts; never hardcoded |
| Quota abuse | Redis ratelimit gates (see `30_QUOTA_ENFORCEMENT_FLOW1.md`) |
---
## 6·Build & release pipeline (TL;DR)
* **Git commits** trigger CI → unit / integration / E2E tests.
* Successful merge to `main`:
* Build `.NET {{ dotnet }}` trimmed selfcontained binary.
* `docker build --sbom=spdx-json`.
* Sign image and tarball with Cosign.
* Attach SBOM + provenance; push to registry and download portal.
---
## 7·Future extraction path
Although the default deployment is a single container, each subservice can be
extracted:
* Concelier → standalone cron pod.
* Policy Engine → sidecar (OPA) with gRPC contract.
* ResultSink → queue worker (RabbitMQ or Azure Service Bus).
Interfaces are stable **as of v0.2 β**; extraction requires a recompilation
only, not a fork of the core.
---
*Last updated {{ "now" | date: "%Y%m%d" }} constants autoinjected.*
- `docs/07_HIGH_LEVEL_ARCHITECTURE.md`
- `docs/24_OFFLINE_KIT.md`
- `docs/09_API_CLI_REFERENCE.md`
- `docs/modules/platform/architecture-overview.md`