Skip to content

Query Scheduler

Status: 📋 Planned
Target Version: v5.5
Feature ID: F1.12 Enhancement

Overview

Priority-based query scheduling system for optimizing resource utilization and meeting SLAs.

Planned Features

  • Priority Scheduling: Multi-level priority queue (High/Medium/Low)
  • Fair Queuing: Prevent starvation of low-priority queries
  • Preemption: Ability to pause/resume long-running queries
  • SLA-Based Scheduling: Prioritize queries to meet SLA targets
  • Workload-Aware Scheduling: OLTP vs OLAP vs Batch differentiation

Architecture (Planned)

Query Admission → Scheduler → [High/Med/Low Priority Queues]
                       ↓
                  Executor Pool (Dynamic)
                       ↓
                  Resource Monitor

Scheduling Policies (Draft)

  1. FIFO: First-in-first-out (default for same priority)
  2. Shortest Job First: Prioritize fast queries
  3. Fair Share: Equal CPU/memory allocation per tenant
  4. SLA-Driven: Prioritize queries at risk of SLA violation

Configuration (Draft)

query_scheduler:
  policy: sla_driven
  priorities:
    - level: high
      max_concurrent: 100
      timeout_ms: 5000
    - level: medium
      max_concurrent: 200
      timeout_ms: 30000
    - level: low
      max_concurrent: 50
      timeout_ms: 300000

Implementation Status

Current: Not implemented
Roadmap: v5.5 Phase 1 (Foundation)
Estimated Effort: 4-5 person-weeks

Note: This is a placeholder document. Implementation details will be added during v5.5 development.

For current query optimization features, see: - Workload-Aware Query Optimization - Distributed Query Optimization