Skip to content

update roadmap current state and phase 1 #141

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Jovonni wants to merge 1 commit into
base: master
Choose a base branch
from
Open

update roadmap current state and phase 1 #141

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 3 additions & 6 deletions ROADMAP.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,9 +24,9 @@ OpenUBA aims to be the standard open-source User & Entity Behavior Analytics (UE

### Model Registry & Ecosystem
- Multi-backend model registry with adapter pattern (`core/registry/registry_service.py`)
- 11 functional adapters split into code registries (local FS, GitHub, OpenUBA Hub) and weights registries (local FS, HuggingFace, Kubeflow), plus 5 legacy generic adapters
- Adapters for code registries (local FS, GitHub, OpenUBA Hub) and weights registries (local FS, HuggingFace, Kubeflow)
- Registry service with unit tests (`core/tests/test_registry/`)
- SHA-256 hashing scaffold (`core/hash.py`) — not yet wired to install-time integrity verification
- Install-time SHA-256 model integrity verification (`core/services/model_installer.py`) that gates installation on a checksum match
- **Community model marketplace — OpenUBA Hub LIVE at https://openuba.org** (Next.js Model Hub in sibling repo `openuba-model-hub`, CNAME `openuba.org`, static catalog of reference models). Backend client adapter ships in `core/registry/adapters/openuba_hub_adapter.py`.

### Scheduling & Async
Expand Down Expand Up @@ -67,9 +67,7 @@ OpenUBA aims to be the standard open-source User & Entity Behavior Analytics (UE

Honest list of items the Current State touches but does not fully deliver, with code citations:

- Postgres migration is incomplete — `core/model.py`, `core/user.py`, `core/process.py` still call `ReadJSONFileFS` / `WriteJSONFileFS` on `core/storage/*.json` (`users.json`, `models.json`, `model_sessions.json`, `scheme.json`). Repos (`core/repositories/`) + migration script (`core/db/migrations/migrate_from_files.py`) exist; cutover in progress.
- `/metrics` endpoint at `core/api_routers/data.py:201` emits domain JSON (Spark/ES counters), not Prometheus exposition format. No `opentelemetry-*` or `prometheus_client` in `requirements.txt`; no OpenTelemetry SDK init in `core/`.
- `core/hash.py` `HashLargeFile` references undefined `filename` / `hashlib` and is not wired into any install-time integrity flow. Scaffold only.
- `core/tests/test_graphql.py` is a 24-LOC smoke test (GETs `/`, checks for an `endpoints` key) — does not exercise a GraphQL query, schema introspection, or PostGraphile.
- `core/registry/adapters/openuba_hub_adapter.py` now defaults to `https://openuba.org`, but the live Hub serves a static Next.js catalog rather than the `/ml/` JSON contract the adapter expects — Hub-side JSON endpoint needs publishing.
- `k8s/postgres.yaml` is a vanilla single-replica Deployment + PVC, not CloudNativePG. HA Postgres moved to Phase 1.
Expand All @@ -80,10 +78,8 @@ Honest list of items the Current State touches but does not fully deliver, with
- [ ] Horizontal pod autoscaling for Spark workers (`k8s/spark-deployment.yaml` hardcodes `replicas: 1`; no `autoscaling/v2` resources anywhere)
- [ ] Multi-tenant isolation (namespace-per-tenant + `tenant_id` across tables / repositories / RBAC; today there is one `openuba` namespace and zero tenant-scoped code)
- [ ] Production-grade observability — Prometheus exposition-format `/metrics` + OpenTelemetry SDK self-instrumentation (current: domain-JSON metrics endpoint only)
- [ ] Complete the JSON-file → Postgres cutover (remove `ReadJSONFileFS` / `WriteJSONFileFS` callers in `core/model.py`, `core/user.py`, `core/process.py`; make migration mandatory on startup)
- [ ] Migrate Postgres deployment to CloudNativePG operator (HA `Cluster` CR, automated failover, scheduled backups) — moved here from Current State per audit
- [ ] GraphQL query-level test coverage (replace smoke test with real query / mutation / schema-introspection suite against PostGraphile)
- [ ] Wire SHA-256 model integrity check into install path (fix `core/hash.py:HashLargeFile`, gate `model_installer` on checksum)

## Phase 2: CNCF Integration (Q4 2026)

Expand Down Expand Up @@ -111,3 +107,4 @@ Honest list of items the Current State touches but does not fully deliver, with
### Note on governance maturity (longest pole)

The governance framework is **shipped** — `GOVERNANCE.md`, `MAINTAINERS.md`, `CONTRIBUTING.md`, `CODE_OF_CONDUCT.md`, `SECURITY.md`, `ADOPTERS.md` (PR #137). What is missing is *demonstration of using it*: the 2-maintainer approval path defined in `GOVERNANCE.md` is currently inoperable with 1 maintainer, `ADOPTERS.md` has 1 entry (the host org), there are no public TSC meeting notes, and no governance-tagged decisions on record. Both blockers resolve only via sustained external contributor + adopter outreach — i.e., they bottleneck on Phase 3's contributor-diversity item, which is the steepest CNCF Sandbox → Incubation gate this project faces.

Loading