ai-center

2022.4

true

Getting Started
- About this guide
- About installation
  - Supported use cases for single-node and multi-node installations
  - Downloading installation packages
Network requirements
- Requirements
Single-node requirements and installation
Multi-node requirements and installation
Post-installation
- Overview
Provisioning a GPU
- Provisioning a GPU
Using the configuration file
- About the configuration file
Node scheduling
- Manage node scheduling
Migration and upgrade
- Migrating to the standalone or Automation Suite environment
- Upgrading AI Center
Basic Troubleshooting Guide
- General AI Center troubleshooting and FAQs
- AI Center standalone troubleshooting

AI Center Installation Guide

DELIVERY:

Automation Cloud Automation Suite Standalone

Last updated Jun 6, 2024

About the configuration file

The cluster_config.json file defines the parameters, settings, and preferences applied to the UiPath services. You need to update this file if you want to change defaults and use any advanced configuration for your cluster.

To edit cluster_config.json, you can use either:

a Linux text editor, such as vi or GNU nano, directly on the Linux machine via SSH (e.g., command: vi cluster_config.json);
your favorite text editor and then copy-paste the file on the machine.

To enable or disable a service via the cluster_config.json file, use true or false for the enabled flag.

General configuration

Mandatory parameters	Description
`fqdn`	The load balancer (multi-node) or machine (single-node) domain name.
`fixed_rke_address`	Fixed address used to load balance node registration and kube API requests. This should be fqdn if load balancer is configured as recommended. Otherwise FQDN of 1st Server Node. Refer to Configuring the load balancer. Can be the IP/FQDN of the first rke2 server in your setup.
`multinode`	Set to `true` if going for multi node installation
`admin_username`	The username that you would like to set as admin (such as: admin ) for the host organization.
`admin_password`	The host tenant admin password to be set.
`rke_token`	Use a newly generated GUID here. This is a pre-shared, cluster-specific secret. It is needed for all the nodes joining the cluster.
`profile`	Sets the profile of the installation. The available profiles are: `default`: single-node evaluation profile. `ha`: multi-node HA-ready production profile. The value of this flag is set automatically by the interactive install wizard. It is used for internal purposes only and should not be modified manually.
`gpu_support`	`true` or `false` - enable or disable GPU support for the cluster. Set to 'true' if you have agent nodes with GPUs. By default it is set to 'false'
`infra.docker_registry.username`	The username that you would like to set for the docker registry installation.
`infra.docker_registry.password`	The password that you would like to set for the docker registry installation.

Optional parameters	Description
`telemetry_optout`	`true` or `false` - used to opt-out of sending telemetry back to UiPath. It is set to `false` by default. If you wish to opt out, then set to `true`.

AI Center parameters	Description
`sql_connection_str`	The SQL connection name.
`orchestrator_url`	The Orchestrator URL address.
`identity_server_url`	The Identity Server URL adress.
`orchestrator_cert_file_path`	The absolute path to Orchestrator certificate file.
`identity_cert_file_path`	The absolute path to Identity Server certificate file.
`identity_access_token`	The access token taken from Identity Server.

Setting certificates

Please refer to the prerequisite document to obtain certificate: .

If no certificate is provided at time of installation, the installer creates a self-issued certificate and configures it in the cluster. The validity of the certificate is 90 days.

In multi-node installations, a certificate is required only on the first node.

Note: Make sure to specify the absolute path for certificate files. Run pwd to get the path of the directory where files are placed and append the certificate file name in the cluster_config.json.

In multi-node installations, a certificate is required only on the first node.

Parameter	Description
`server_certificate.ca_cert_file`	Absolute path to Certificate Authority certificate. This certificate is the authority that signs the TLS certificate. In case of self-signed is the `rootCA.crt` created in earlier steps. Leave blank, in case you want installer to generate.
`server_certificate.tls_cert_file`	Absolute path to TLS certificate (`server.crt` for self-signed created in earlier steps). Leave blank, in case you want installer to generate.
`server_certificate.tls_key_file`	Absolute path to certificate key (`server.key` for self-signed created in earlier steps). Leave blank, in case you want installer to generate.
`additional_ca_certs`	Absolute path to the file containing additional CA certificates that you want to be trusted by all the services running as part of Automation Suite. All certificates in the file should be valid `PEM` format. For example, you need to provide the file containing the SQL server CA certificate if the certificate is not issued by a public certificate authority.

Database configuration

Automatically create the necessary databases

If you want the installer to create the databases, then please fill in the following fields

Parameter	Description
`sql.create_db`	Set to `true`.
`sql.server_url`	FQDN of sql server, where you wish to installer to configure database
`sql.port`	port number on which a database instance should be hosted in sql server
`sql.username`	username / userid to connect to sql server.
`sql.password`	password of username provided earlier to connect to sql server.

Note:

Ensure the user has the dbcreator role. This grants them permission to create the database in SQL Server. Otherwise the installation fails.

ODBC connection does not support usernames that contain special characters. For database usernames for AI Center and Document Understanding, use only uppercase and lowercase letters.

Bring your own database

In case if you are providing the database, then we need SQLconnection strings for every database. The following SQL connection string formats are supported.

SQL connection string formatt	Products
JDBC	AI Center
ODBC	Document Understanding

Parameter	Description
`sql_connection_string_template`	Full ADO.NET connection string where Catalog name is set to DB_NAME_PLACEHOLDER. The installer will replace this placeholder with the default database names for the installed suite services.
`sql_connection_string_template_jdbc`	Full JDBC connection string where database name is set to DB_NAME_PLACEHOLDER. The installer will replace this placeholder with the default database names for the installed suite services.
`sql_connection_string_template_odbc`	Full ODBC connection string where database name is set to DB_NAME_PLACEHOLDER. The installer will replace this placeholder with the default database names for the installed suite services. This parameter is used by Document Understanding.

Important:

If you manually set the connection strings in the configuration file, you can escape SQL, JDBC, or ODBC passwords as follows:

for SQL: add ' at the beginning and end of the password, and double any other '.
for JDBC/ODBC: add { at the beginning of the password and } at the end, and double any other }.

ql_connection_string_template example

Server=tcp:sfdev1804627-c83f074b-sql.database.windows.net:1433;Initial Catalog=DB_NAME_PLACEHOLDER;Persist Security Info=False;User Id=testadmin@sfdev1804627-c83f074b-sql.database.windows.net;Password=***;MultipleActiveResultSets=False;Encrypt=True;TrustServerCertificate=False;Connection Timeout=30;Max Pool Size=100;Server=tcp:sfdev1804627-c83f074b-sql.database.windows.net:1433;Initial Catalog=DB_NAME_PLACEHOLDER;Persist Security Info=False;User [email protected];Password=***;MultipleActiveResultSets=False;Encrypt=True;TrustServerCertificate=False;Connection Timeout=30;Max Pool Size=100;

ql_connection_string_template_jdbc example

jdbc:sqlserver://sfdev1804627-c83f074b-sql.database.windows.net:1433;database=DB_NAME_PLACEHOLDER;user=testadmin;password=***;encrypt=true;trustServerCertificate=false;Connection Timeout=30;hostNameInCertificate=sfdev1804627-c83f074b-sql.database.windows.net"jdbc:sqlserver://sfdev1804627-c83f074b-sql.database.windows.net:1433;database=DB_NAME_PLACEHOLDER;user=testadmin;password=***;encrypt=true;trustServerCertificate=false;Connection Timeout=30;hostNameInCertificate=sfdev1804627-c83f074b-sql.database.windows.net"

sql_connection_string_template_odbc example</summary>

SERVER=sfdev1804627-c83f074b-sql.database.windows.net,1433;DATABASE=DB_NAME_PLACEHOLDER;DRIVER={ODBC Driver 17 for SQL Server};UID=testadmin;PWD=***;MultipleActiveResultSets=False;Encrypt=YES;TrustServerCertificate=NO;Connection Timeout=30;"SERVER=sfdev1804627-c83f074b-sql.database.windows.net,1433;DATABASE=DB_NAME_PLACEHOLDER;DRIVER={ODBC Driver 17 for SQL Server};UID=testadmin;PWD=***;MultipleActiveResultSets=False;Encrypt=YES;TrustServerCertificate=NO;Connection Timeout=30;"

For AI Center, set the following aicenter SQL configuration parameter:

Parameter	Description
`aicenter.ai_appmanager.sql_connection_str`	AI app manager JDBC connection string (Refer below for the JDBC format)

Note: Check the JDBC connection string has the format in the sample below.

Sample AI Center connection string

"aicenter": {
    "enabled": true,
    "sql_connection_str": "jdbc:sqlserver://sfdev1804627-c83f074b-sql.database.windows.net;instanceName=instance;database=aicenter;[email protected];password=TFgID_9GsE7_P@srCQp0WemXX_euHQZJ"
  }"aicenter": {
    "enabled": true,
    "sql_connection_str": "jdbc:sqlserver://sfdev1804627-c83f074b-sql.database.windows.net;instanceName=instance;database=aicenter;[email protected];password=TFgID_9GsE7_P@srCQp0WemXX_euHQZJ"
  }

Sample DU connection string

Note: Only use this if you want to replace the default database with your own.

"documentunderstanding":
{
  "datamanager": {
    "sql_connection_str": "mssql+pyodbc://testadmin:[email protected]:1433/datamanager?driver=ODBC+Driver+17+for+SQL+Server",
  }"documentunderstanding":
{
  "datamanager": {
    "sql_connection_str": "mssql+pyodbc://testadmin:[email protected]:1433/datamanager?driver=ODBC+Driver+17+for+SQL+Server",
  }

Important: Make sure the SQL account specified in the connection strings is granted the db_owner role for all databases. If security restrictions do not allow the use of db_owner, then the SQL account should have the following roles and permissions on all databases:

db_ddladmin
db_datawriter
db_datareader
EXECUTE permission on dbo schema

Monitoring configuration

To provision enough resources for monitoring (see Using the monitoring stack), you should consider the number of vCPUs in the cluster and the amount of desired metric retention. See below for how to set the following monitoring resource configurations.

The following table describes the monitoring field details:

Parameter	Description
`prometheus_retention`	In days. This is the amount of days that metrics will be retained for the purpose of visualization in Grafana and manual querying via the Prometheus console. Default value is `7`.
`prometheus_storage_size`	In GB. Amount of storage space to reserve per Prometheus replica. A good rule of thumb is to set this value to: `0.65 * vCPU cores * (prometheus_retention / 7)` Example: If you set `prometheus_retention` to 14 days, and your cluster has 80 cores spread across 5 machines, this becomes: `0.65 * 80 * (14 / 7)` `52 * (2)` `104` Default value is `45` and should not be set lower. If Prometheus starts to run out of storage space, an alert will be triggered with specific remediation instructions.
`prometheus_memory_limit`	In MB. Amount of memory to limit each Prometheus replica to. A good rule of thumb is to set this value to: `100 * vCPU cores * (prometheus_retention / 7)` Example: If you've set `prometheus_retention` to 14 days, and your cluster has 80 cores spread across 5 machines, this becomes: `100 * 80 * (14 / 7)` `8000 * (2)` `16000` Default value is `3200` for the single-node evaluation mode, and `6000` for the multi-node HA-ready production mode, and should not be set lower. If Prometheus starts to run out of memory, an alert will be triggered with specific remediation instructions. See here.

Example:

"monitoring": {
  "prometheus_retention": 14,
  "prometheus_memory_limit": 16000,
  "prometheus_storage_size": 104
}"monitoring": {
  "prometheus_retention": 14,
  "prometheus_memory_limit": 16000,
  "prometheus_storage_size": 104
}

Optional: Adding proxy configuration

Note: Make sure you meet the proxy server requirements before configuring the proxy server during installation.

While running the interactive installer wizard, you need to exit it and update the cluster_config.json during the advanced configuration step.

You need to add the following to the configuration file using vim or your favorite editor to the configuration file:

"proxy": {
      "enabled": "true",
      "http_proxy": "http://<PROXY-SERVER-IP>:<PROXY-PORT>",
      "https_proxy": "http://<PROXY-SERVER-IP>:<PROXY-PORT>",
      "no_proxy": "<Comma separated list of ips that should not got though proxy server>"
    }"proxy": {
      "enabled": "true",
      "http_proxy": "http://<PROXY-SERVER-IP>:<PROXY-PORT>",
      "https_proxy": "http://<PROXY-SERVER-IP>:<PROXY-PORT>",
      "no_proxy": "<Comma separated list of ips that should not got though proxy server>"
    }

Allow 30070 port on the VM for inbound and outbound calls.

Mandatory parameters	Description
`enabled`	Use `true` or `false` to enable or disable proxy settings.
`http_proxy`	Used to route HTTP outbound requests from the cluster. This should be the proxy server FQDN and port.
`https_proxy`	Used to route HTTPS outbound requests from the cluster. This should be the proxy server FQDN and port.
`no_proxy`	Comma-separated list of hosts, IP addresses, or IP ranges in CIDR format that you do not want to route via the proxy server. This should be a private subnet range, sql server host, named server address, metadata server address: `*.<fqdn>,<fixed_rke_address>:9345,<fixed_rke2_address>:6443`. `fqdn` - the cluster FQDN defined in `cluster_config.json` `fixed_rke_address` - the `fixed_rke_address` defined in `cluster_config.json` `named server address` - the IP address from `/etc/resolv.conf` `private_subnet_ip` - the cluster VNet `sql server host` - sql server host `metadata server address` - the IP address `169.254.169.254` used to fetch machine metadata by cloud services such as Azure and AWS

Optional: Enabling resilience to zonal failures in a multi-node HA-ready production cluster

To enable resilience to zonal failures in a multi-node cluster, take the following steps:

Make sure nodes are spread evenly across three availability zones. For a bare-metal server or VM provided by any vendor except for AWS, Azure, or GCP, zone metadata has to be provided via the configuration file at /etc/default/k8s-node-labels on every machine in following format.

NODE_REGION_LABEL=<REGION_NAME>
NODE_ZONE_LABEL=<ZONE_NAME>
cat > /etc/default/k8s-node-labels <<EOF
EXTRA_K8S_NODE_LABELS="topology.kubernetes.io/region=$NODE_REGION_LABEL,topology.kubernetes.io/zone=${NODE_ZONE_LABEL}"
EOFNODE_REGION_LABEL=<REGION_NAME>
NODE_ZONE_LABEL=<ZONE_NAME>
cat > /etc/default/k8s-node-labels <<EOF
EXTRA_K8S_NODE_LABELS="topology.kubernetes.io/region=$NODE_REGION_LABEL,topology.kubernetes.io/zone=${NODE_ZONE_LABEL}"
EOF

Update the cluster_config.json file during the advanced configuration step.

To update the cluster_config.json using the interactive installation wizard, exit at advanced configuration step and add the following to the configuration file using vim or your favorite editor:

"zone_resilience": true"zone_resilience": true

Mandatory parameters	Description
`zone_resilience`	Use `true` or `false` to enable or disable resilience to zonal failure.

Optional: Passing custom Resolv.conf

The Kubernetes cluster that Automation Suite deploys uses the name servers configured in /etc/resolv.conf. Kubernetes does not work with local DNS resolvers (127.0.0.1 or 127.0.0.0/8), so if you have such name servers configured in /etc/resolv.conf file, you need to pass a file reference with the correct nameserver entries accessible from anywhere on the VM in the .infra.custom_dns_resolver parameter in cluster_config.json.

For details on a known limitation, see Kubernetes documentation.

Optional parameter	Description
`.infra.custom_dns_resolver`	Path to the file with correct name server entries that can be accessed from anywhere on the VM. These nameserver entries must not be from `127.0.0.0/8`.

On this page

General configuration
Setting certificates
Database configuration
Automatically create the necessary databases
Bring your own database
Monitoring configuration
Optional: Adding proxy configuration
Optional: Enabling resilience to zonal failures in a multi-node HA-ready production cluster
Optional: Passing custom Resolv.conf

Was this page helpful?

PREVIOUSProvisioning a GPU

NEXTManage node scheduling

Support and Services

Get The Help You Need

UiPath Academy

Learning RPA - Automation Courses

UiPath Forum

UiPath Community Forum

Trust and Security

Cookies Policy