AWS deadline-cloud documentation change
Summary
Added GPU utilization monitoring details to worker dashboard documentation and fixed punctuation
Security assessment
Changes focus on adding GPU monitoring capabilities and improving grammar. No security vulnerabilities, configurations, or incidents are mentioned. The additions are purely performance monitoring enhancements.
Diff
diff --git a/deadline-cloud/latest/userguide/view-worker-dashboard.md b/deadline-cloud/latest/userguide/view-worker-dashboard.md index 6e6614baa..ddfb5ce65 100644 --- a//deadline-cloud/latest/userguide/view-worker-dashboard.md +++ b//deadline-cloud/latest/userguide/view-worker-dashboard.md @@ -13 +13 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * Metadata, such as the instance type, for the worker + * Metadata, such as the instance type, for the worker. @@ -15 +15 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * The session actions that the worker performed + * The session actions that the worker performed. @@ -17 +17 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * Worker performance, including CPU, memory and disk usage + * Worker performance, including CPU, memory, and disk usage. The dashboard also shows GPU utilization for GPU-accelerated instances. @@ -19 +19 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * A graph of the CPU, memory, and disk usage over time + * A graph of the CPU, memory, and disk usage over time. The graph also includes GPU utilization for GPU-accelerated instances. @@ -21 +21 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * A graph of the disk speed over time + * A graph of the disk speed over time. @@ -23 +23 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo - * The worker log for the task + * The worker log for the task. @@ -58 +58 @@ The _worker dashboard_ provides details for the worker that processes a task. Yo -When renders take longer than expected, the worker dashboard can help determine if your instances are adequately sized for your workloads. While 100% vCPU utilization is normal for many renderers, consistently high memory usage near maximum capacity and elevated disk space utilization may indicate that your instances are under-provisioned. In such cases, upgrading your fleet's instance configuration can reduce render errors and significantly improve render times. However, it's important to continue monitoring worker performance after upgrading to ensure you've found the optimal balance - upgrading too aggressively can lead to unnecessary costs through over-provisioning. +When renders take longer than expected, the worker dashboard can help determine if your instances are adequately sized for your workloads. While 100% vCPU utilization is normal for many renderers, consistently high memory usage near maximum capacity and elevated disk space utilization may indicate that your instances are under-provisioned. For GPU-accelerated workloads, consistently high GPU utilization might also indicate that you need additional GPU capacity. In such cases, upgrading your fleet's instance configuration can reduce render errors and significantly improve render times. However, it's important to continue monitoring worker performance after upgrading to ensure you've found the optimal balance - upgrading too aggressively can lead to unnecessary costs through over-provisioning.