Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Batch capacity refers to the maximum number of batch tasks that can be processed at a time. It depends on both the number of batch servers and the number of batch threads available for processing these tasks.
To calculate the batch capacity, multiply the number of batch servers by the number of batch threads per server:
Batch capacity = Number of batch servers × Number of batch threads per server
The total batch capacity for the environment is determined based on user licenses. We establish the minimum and maximum number of batch servers required to serve this batch capacity.
To view batch capacity, use System Administration > Setup > Server configuration and look for available batch servers.
Batch auto scaling
Auto scaling is a new feature that automatically adjusts your batch servers according to resource usage thresholds. It provides elasticity to your environment, allowing it to adapt to varying workloads dynamically. This process is entirely automated and relies on predefined signals based on CPU and memory usage of batch servers.
Auto scaling becomes beneficial when the workload on an environment fluctuates over time. We continuously monitor the reported load and periodically evaluate triggers to determine if scaling is necessary.
The lower load threshold signifies the point at which the service scales in. If the average load falls below this threshold, the service scales in.
Conversely, the upper load threshold indicates when the service scales out. If the average load exceeds this threshold, the service scales out.
Note
- For batch auto scaling to work, your environment should have batch priority-based scheduling enabled, and your PU should be 10.0.26 (PU 50) or higher.
- After batch auto scaling is activated for the environment, the platform periodically adjusts the thread count for each server as per batch capacity. Any manual alterations to the thread count are disregarded and overridden by the platform's automated processes.
In PBS-enabled environments, autoscaling now ensures a constant total thread count across all batch servers, regardless of scale-up or scale-down events.
For example, your environment starts with six batch servers, each configured with eight threads, totaling 48 threads. As the environment scales, let's say if autoscaling increases the number of servers to 8, the system automatically adjusts each server to use six threads, keeping the total at 48 (8 × 6 = 48). Similarly, if the environment scales down to four servers, each one is assigned 12 threads to maintain the same total (4 × 12 = 48).
In both cases, the overall thread capacity stays consistent—only the distribution changes.
This approach ensures that batch processing capacity remains consistent, even when CPU or memory usage alone wouldn't trigger autoscale.
Note
- The thread count per server doesn't exceed 16, in line with platform safeguards to prevent resource saturation. By default, the thread count is set to 8. When autoscaling is enabled, users can no longer manually configure the thread count, as it's automatically managed by the platform.
How to increase batch capacity
To increase the batch capacity in a production environment, you must acquire more user licenses and update subscription estimates in Microsoft Dynamics Lifecycle Services. For updated user licenses, we automatically increase the batch capacity by adjusting the thread count per existing server. The platform adds more batch servers after the existing batch servers reach their threshold limits for CPU and memory utilization.
To increase the batch capacity in a sandbox environment, you need a Tier-4 or Tier-5 sandbox. This action isn't possible in Tier-2 or Tier-3 sandboxes.
For more information about capacity planning, see Environment planning.