Back to All Scenarios
PASSEDdatabase / gc_pressure
ScyllaDB 6.0 GC Pauses Causing Latency Spikes on CentOS 9 Stream
ScyllaDB 6.0 is experiencing excessive garbage collection pauses. Response times spiking during GC cycles. P99 latency 10x normal.
Pattern
PERFORMANCE_DEGRADATION
Severity
MEDIUM
Confidence
51%
Remediation
Auto-Heal
Test Results
| Metric | Expected | Actual | Result |
|---|---|---|---|
| Pattern Recognition | PERFORMANCE_DEGRADATION | PERFORMANCE_DEGRADATION | |
| Severity Assessment | HIGH | MEDIUM | |
| Incident Correlation | N/A | None | |
| Cascade Escalation | N/A | No | |
| Remediation | — | Auto-Heal — Corax resolves autonomously |
Scenario Conditions
CentOS 9 Stream. ScyllaDB 6.0 GC pause time 350ms. GC frequency: every 5 seconds. Heap at 90%. Response time degraded 48x.
Injected Error Messages (1)
ScyllaDB 6.0 degraded performance on CentOS 9 Stream — garbage collection pauses 350ms, response time spike 48x normal, high latency during GC cycles, P99 latency critical
Neural Engine Root Cause Analysis
Performance degradation detected — response times are significantly above baseline, indicating resource pressure, increased load, or infrastructure issues.
Remediation Plan
Investigate server resource usage (CPU, memory, disk I/O). Check for background jobs or traffic spikes. Optimize slow queries or endpoints if specific to one service.
Improvements Applied
- Severity: MEDIUM (expected HIGH)
Tested: 2026-04-02Monitors: 1 | Incidents: 1Test ID: cmnhnvz490352lijgfrb83z87