Inference AvailabilitySLO MET
99.98%
Target: 99.9% | Past 30 Days
Inference Latency (p95)FAST
284 ms
TTFT p50: 142ms | TPS: 38.6 t/s
Error BudgetHEALTHY
94.2%
Burn Rate: 0.08x | Impact Budget Ok
Queue BackpressureNOMINAL
0 / 50
Concurrency Slot: 1/1 Active (Fairness On)
OCI A1 ARM Host TelemetryNODE ONLINE
CPU (2 OCPU Ampere)
18.4%
Load: 0.34, 0.42, 0.38
Memory (16 GB Total)
5.8 GB
Utilization: 36.2% | Headroom Ok
Swap Storage
0 MB / 4 GB
Zero Swap Thrashing
Locked AI Model
Qwen3-8B
5.03 GB GGUF | Thinking Mode: Off
Contextual Operational WorklistNORMAL
Active Principal Role: Platform Superadmin (Tier 1)
Active Platform Mode: NORMAL
Pending Durable Approvals: 0 items awaiting decision
Active Scheduled Changes: Zero active maintenance windows