Adaptix Insight · Executive workspace
Live executive surface
Infrastructure compute control
AI Compute Control
AI infrastructure telemetry unavailableThis page previously displayed fabricated GPU utilization, memory, thermal, queue, and model inventory numbers. Until the backend emits AI-specific infrastructure telemetry, it now stays explicit about what is not instrumented.
No GPU metrics endpointNo inference queue APINo model registry telemetry
GPU utilization
Unavailable
No founder endpoint currently reports AI accelerator utilization.
Memory in use
Unavailable
No backend source emits resident model or KV-cache memory figures.
Temperature
Unavailable
Thermal telemetry is not exposed to the founder infrastructure surface.
Power draw
Unavailable
No power telemetry API exists for AI compute in this repository.
Active jobs
Unavailable
Inference queue state is not currently emitted by a backend service.
Queue depth
Unavailable
Pending AI workload backlog is not instrumented.
Inference queue
Active AI jobs
The repository does not currently expose a live inference-queue endpoint for founder operations.
Queue telemetry
No backend API currently returns running or queued AI jobs for this dashboard.
Unavailable
Operator action
This page now refuses to invent job IDs, durations, or statuses without a real source.
Block fabricated metrics
Inference posture
Model latency overview
Latency and model-health posture require real inference observability that is not yet present.
Per-model latency
No AI metrics endpoint publishes latency by model or workflow.
Unavailable
Throughput posture
Requests-per-minute and token-throughput are not instrumented in founder APIs.
Unavailable
Health posture
No backend service currently publishes stable/watch/healthy states for model runtime.
Unavailable
Memory residency
GPU memory allocation
This breakdown is intentionally empty until a real AI memory telemetry source exists.
Weights
Unavailable
KV cache
Unavailable
Execution buffers
Unavailable
Available headroom
Unavailable
Thermal envelope
Temperature and throughput
AI-specific thermal and throughput telemetry is not yet surfaced to this dashboard.
Current temperature
Unavailable
1h thermal average
Unavailable
Tokens / sec
Generation throughput is not emitted by any current founder API.
Unavailable
Requests / min
Aggregate AI ingress is not exposed to this control surface.
Unavailable
Batch efficiency
Scheduler efficiency is not currently measured here.
Unavailable
Model inventory
Model registry
Versioned AI model inventory is not yet backed by a model-registry API in this repository.
Registry telemetry
No founder endpoint currently returns model names, versions, sizes, or update timestamps.
Unavailable
Why this changed
Fabricated model inventory and GPU readings were removed to avoid false operational confidence.
Truth-first