Autonomous Maintenance

How the SveltyCMS System Watchdog performs proactive self-healing and performance optimization.

4/11/2026

2 min read

On this page

SveltyCMS 2026 introduces Autonomous Maintenance, a background self-healing layer that proactively optimizes the system without human intervention. This is orchestrated by the System Watchdog.

🐕 The System Watchdog

The Watchdog is a persistent background service initialized during the WARMING phase. It executes a health check loop every 5 minutes.

Core Responsibilities

Service Re-initialization: If a critical adapter (e.g., Database or Auth) reports a FAILED state, the Watchdog attempts a surgical reinitializeSystem() call.
Resource Monitoring: Tracks RAM and CPU spikes, logging anomalies to the Enterprise Monitor.
Drift Detection: Verifies that the database schema remains in sync with the filesystem-based “Source of Truth.”

⚡ Cache Compaction

To prevent memory bloat in high-traffic environments, the Watchdog monitors the internal L1 Cache (in-memory store).

Threshold: Compaction is triggered when the cache exceeds 10,000 items or a 30% fragmentation level.
Action: The system performs a “LRU Sweep,” purging the least-recently-used items across all tenants while preserving “Hot” system settings and auth sessions.
Result: Ensures the SvelteKit server maintains a stable memory footprint, even during massive content imports.

🗄️ Database Index Optimization

The Watchdog analyzes query performance bubbled up through the DatabaseResilience micro-telemetry.

Slow Query Detection: If queries consistently exceed 50ms, the Watchdog logs a performance incident.
Auto-Indexing: For MongoDB and SQL adapters, the system can autonomously rebuild missing indexes for fields marked as searchable or unique in the collection schema if it detects a full table scan is occurring.

🗑️ Automated Trash Purge

While SveltyCMS uses “Soft Deletes” by default, the maintenance layer manages the archival lifecycle.

Retention Policy: By default, items in the Trash are kept for 30 days.
Autonomous Purge: During the nightly maintenance window, the Watchdog permanently removes items that have exceeded the retention threshold, reclaiming database storage.

📊 Monitoring

Admins can track autonomous activity in the Enterprise Monitor under the System Vitals quadrant.

Last Run: Timestamp of the most recent maintenance cycle.
Heals Performed: Count of automated service recoveries.
Purge Stats: Number of items cleared from trash/cache.

Documentation