# StellaOps Scheduler

Scheduler detects advisory/VEX deltas, computes impact windows, and orchestrates re-evaluations across Scanner and Policy Engine.

## Responsibilities
- Maintain impact cursors and queues for re-scan/re-evaluate jobs.
- Expose APIs for policy-triggered rechecks and runtime hooks.
- Emit DSSE-backed completion events for downstream consumers (UI, Notify).
- Provide SLA-aware retry logic with deterministic evaluation windows.

## Key components
- `StellaOps.Scheduler.WebService` control plane.
- `StellaOps.Scheduler.Worker` job executor.
- Shared libraries under `StellaOps.Scheduler.*`.

## Integrations & dependencies
- PostgreSQL (schema `scheduler`) for impact models.
- Valkey/NATS for queueing.
- Policy Engine, Scanner, Notify.

## Operational notes
- Monitoring assets in ./operations/worker-grafana-dashboard.json & worker-prometheus-rules.yaml.
- Operational runbook ./operations/worker.md.

## Related resources
- ./operations/worker.md
- ./operations/worker-grafana-dashboard.json
- ./operations/worker-prometheus-rules.yaml

## Backlog references
- SCHED-MODELS-20-001 (policy run DTOs) and related tasks in ../../TASKS.md.
- Scheduler observability follow-ups in src/Scheduler/**/TASKS.md.

## Implementation Status

### Current Objectives
- Maintain deterministic behaviour and offline parity across releases
- Keep documentation, telemetry, and runbooks aligned with latest sprint outcomes
- Coordinate with Policy Engine for incremental re-evaluation workflows

### Epic Milestones
- Epic 2 – Policy Engine & Editor: incremental policy run orchestration, change streams, explain trace propagation (in progress)
- Epic 6 – Vulnerability Explorer: findings updates and remediation triggers integration (in progress)
- Epic 9 – Orchestrator Dashboard: job telemetry and control surfaces for UI/CLI (planned)

### Core Capabilities
- Impact cursor maintenance and queue management for re-scan/re-evaluate jobs
- Change-stream detection for advisory/VEX/SBOM deltas
- Policy-triggered recheck orchestration with runtime hooks
- SLA-aware retry logic with deterministic evaluation windows
- DSSE-backed completion events for downstream consumers

### Integration Points
- PostgreSQL schema (scheduler) for impact models and job state
- Valkey/NATS for queueing with idempotency
- Policy Engine, Scanner, Notify for job coordination
- Orchestrator for backfills and incident routing

### Operational Assets
- Monitoring: worker-grafana-dashboard.json, worker-prometheus-rules.yaml
- Runbooks: operations/worker.md
- Observability: metrics, traces, structured logs with correlation IDs

### Technical Notes
- Coordination approach: review AGENTS.md, sync via docs/implplan/SPRINT_*.md
- Backlog tracking: SCHED-MODELS-20-001 and related tasks in ../../TASKS.md
- Module tasks: src/Scheduler/**/TASKS.md

## Epic alignment
- **Epic 2 – Policy Engine & Editor:** orchestrate incremental re-evaluation and simulation runs when raw facts or policies change.
- **Epic 6 – Vulnerability Explorer:** feed triage workflows with up-to-date job status, explain traces, and ledger hooks.
- **Epic 9 – Orchestrator Dashboard:** expose job telemetry, throttling, and replay controls through orchestration dashboards.