Compute Fabric
Compute Fabric shows CPU and memory capacity across execution hosts together with active and failed VM placement and execution jobs.
Execution capacity and jobs
Compute Fabric summarizes execution hosts, available CPU and memory capacity, active workloads, and VM placement/execution job state across the Continuum.
When a VM fails to start or run, open its job details here before changing the workload. The job record is the most direct view of the execution failure reported by the worker.
Interface tour
Compute Fabric Overview
The overview summarizes the entire STRATUM datacenter: execution hosts, total CPU and memory capacity, running workload jobs, and current utilization across the Continuum.
Compute Job Failure Details
Open a failed job to see the worker, execution context, and reported failure details before retrying or changing the VM.
Compute Fabric Memory Deduplication
STRATUM continuously measures memory deduplication so assigned VM memory does not have to equal physical RAM consumed. Identical guest pages—kernel code, shared libraries, zero-filled memory, common services, and repeated boot state—can be consolidated into shared copy-on-write pages while the VMs keep running. In highly homogeneous fleets, that can reclaim enormous amounts of RAM and can cut physical memory consumption by half or more without reducing the memory configured for the workloads. The RAM-saved value shows the capacity STRATUM is reclaiming from the hardware you already own.
Memory deduplication turns repeated guest memory into measurable host capacity. STRATUM shows the savings live instead of hiding them behind an abstract overcommit ratio.
The outcome you should see
Compute Fabric shows the expected execution capacity and job state. Failed jobs expose enough detail to identify the worker-side execution error before changing the workload.