- Overview
- Requirements
- Pre-installation
- Installation
- Post-installation
- Migration and upgrade
- Upgrading Automation Suite
- Migrating standalone products to Automation Suite
- Step 1: Restoring the standalone product database
- Step 2: Updating the schema of the restored product database
- Step 3: Moving the Identity organization data from standalone to Automation Suite
- Step 4: Backing up the platform database in Automation Suite
- Step 5: Merging organizations in Automation Suite
- Step 6: Updating the migrated product connection strings
- Step 7: Migrating standalone Orchestrator
- Step 8: Migrating standalone Insights
- Step 9: Migrating standalone Test Manager
- Step 10: Deleting the default tenant
- Performing a single tenant migration
- Migrating between Automation Suite clusters
- Migrating from Automation Suite on EKS/AKS to Automation Suite on OpenShift
- Monitoring and alerting
- Cluster administration
- Product-specific configuration
- Accessing the Models
- Enabling or disabling Extended Languages OCR
- Allocating GPU resources for Document Understanding modern projects
- Orchestrator advanced configuration
- Configuring Orchestrator parameters
- Configuring appSettings
- Configuring the maximum request size
- Overriding cluster-level storage configuration
- Configuring NLog
- Saving robot logs to Elasticsearch
- Configuring credential stores
- Configuring encryption key per tenant
- Cleaning up the Orchestrator database
- Skipping host library creation
- Troubleshooting
- The backup setup does not work due to a failure to connect to Azure Government
- Pods in the uipath namespace stuck when enabling custom node taints
- Unable to launch Automation Hub and Apps with proxy setup
- Robot cannot connect to an Automation Suite Orchestrator instance
- Log streaming does not work in proxy setups
- Velero backup fails with FailedValidation error
- Accessing FQDN returns RBAC: access denied error
Automation Suite on EKS/AKS installation guide
Allocating GPU resources for Document Understanding modern projects
For optimal performance, allocating the appropriate number of GPU resources is vital. You allocate GPU resources through ArgoCD, which contains specific parameters that correspond to each service. By adjusting these parameters, you can control the distribution of GPU resources as per the needs of each service.
Each service listed in the capacity calculator has a corresponding ArgoCD parameter as follows:
- OCR replicas corresponds to the
du-ocr-saas.replicaCountparameter. - Custom model training replicas corresponds to the
du-training-ssde.replicasparameter. - Custom model replicas corresponds to the
du-aimodelhost-2404.replicasparameter. - OOB model replicas corresponds to the
du-ssde-saas.replicaCountparameter.
Before allocating of GPU resources to each service, use the capacity calculator to determine the appropriate number of GPUs that should be allocated to each service. The output is based on your specific needs and will serve as a guide when updating the parameters in ArgoCD.
For more information, check the UiPath Automation Suite Install Sizing Calculator page.
For Document Understanding modern projects, the minimum recommended GPU is NVIDIA T4.
-
Open ArgoCD.
-
Select the documentunderstanding application.
-
Select Details to open the Details menu.
-
Select the Parameters tab.
-
Select Edit so you can edit the parameter values.
-
Search for the parameters that need to be edited and allocate the needed number of GPUs.
Each service listed in the capacity calculator has its corresponding ArgoCD parameter as follows:
- OCR replicas corresponds to the
du-ocr-saas.replicaCountparameter. - Custom model training replicas corresponds to the
du-training-ssde.replicasparameter. - Custom model replicas corresponds to the
du-aimodelhost-2404.replicasparameter. - OOB model replicas corresponds to the
du-ssde-saas.replicaCountparameter.
- OCR replicas corresponds to the
-
Select Save.