Hardening and scaling
This page collects what you need to run InfoConnect Hub in production: security hardening, authentication and access control, availability and scaling, and monitoring.
You can configure most of the Hub through the Admin API. A few hardening steps must be done directly on the Hub instance. After the Hub is configured, it communicates with external services and IBM i through the Functional API endpoints. For the full list of endpoints, see the API Reference.
Security hardening
By default, the Hub has predefined credentials, an HTTP listener and no IBM i connection. The default Basic Auth credentials are:
- Admin API: user ID
Admin, passwordPassword. - Functional API: user ID
User, passwordPassword@123.
If you start the Hub with the distribution start scripts, the user IDs and passwords are the values that you enter in the first-launch credentials wizard. They are stored in start-scripts/config/.env.
To authenticate through an OAuth provider instead of Basic Auth, use the authorization server. See OAuth Configuration Guide.
An evaluation setup usually has only basic security settings and a non-secure IBM i connection. Complete this checklist early in the evaluation:
- Configure HTTPS for the API / HTTP listener (see Setting up HTTPS).
- Configure a TLS connection to IBM i.
- Change the encryption key for sensitive properties (see Changing the encryption key).
- Encrypt all credentials and other sensitive API properties with the Admin encryption API.
- Change the default user IDs and passwords of the Admin API and the Functional API (see Changing the default user ID and password).
- Change the MCP credentials
mcpUser/mcpUserPassword(mcp/mcpby default) in the MCP and UI configurations, and the users and client secret of the authorization server if you use it. - Restrict access to the APIs to allowed source IP addresses, and add rate limiting and other security policies, at the network, load balancer or API gateway level. The Hub has no built-in IP allow-list; the RPA Hub has one.
In the following sections, <project-folder> is the Hub module folder that contains the config and executable-jars folders. It is <hub-home>/info-connect-hub in the distribution archive and /opt/info-connect-hub on an AWS Marketplace instance.
Setting up HTTPS
To enable HTTPS, create a keystore with a certificate and configure the Hub to use it. This example uses a self-signed certificate.
-
Generate a self-signed certificate with the Java
keytoolcommand. The command asks several questions and creates the keystore filessl-server.jksin the directory where you run it: -
Optional: list the content of the keystore:
-
Copy the keystore file into
<project-folder>/config. -
Add the following entries to
<project-folder>/config/application.properties:server.ssl.key-alias=selfsigned_localhost_sslserver server.ssl.key-store-password=changeit server.ssl.key-store=file:./info-connect-hub/config/ssl-server.jks server.ssl.key-store-provider=SUN server.ssl.key-store-type=JKSThe start scripts run the modules from
<hub-home>, so give the key store path relative to<hub-home>or as an absolute path.To store the keystore password in encrypted form, put the encrypted value in
ENC(): -
Update the modules that call the Hub. In the UI configuration, set
infoconnect.hub.secure=true,infoconnect.hub.truststore.pathandinfoconnect.hub.truststore.pass. In the MCP configuration, sethub.urlto the HTTPS URL of the Hub. -
Restart the Hub.
Encrypting secrets
Passwords, keys and other secrets in connection and Data Flow definitions must be encrypted values. Encrypt them with the Admin API of the Hub. The call needs the credentials of an administrator.
-
Send the plain values in one request. Each key of the JSON object is a name of your choice:
-
The response contains the encrypted values under the same keys:
-
Use the encrypted value in the definition, for example in
passwordof a connection or in the secret fields ofsinkTypePropertiesof a Data Flow. Paste it as it is, withoutENC(...). Inapplication.properties, wrap it inENC(...)instead, see Changing the default user ID and password.
User names and other values that are not secret stay plain. The Hub encrypts with its encryption key, so the values work only on a Hub with the same key. If you change the key, encrypt the secrets again. The RPA Hub has its own encryption endpoint, see RPA Hub user guide.
In the web UI you do not encrypt the values yourself on the Form tab. On the JSON tab, the secrets must already be encrypted.
Changing the encryption key
Before you change the encryption key, make sure that no active connection, poller or encrypted property still uses the old key.
How you change the key depends on how you run the Hub:
- Distribution start scripts (
infoconnect-hub.batorinfoconnect-hub.sh): the credentials wizard asks for the encryption word on the first launch and stores it asENCRYPTION_WORDinstart-scripts/config/.env. To change it later, run the credential update script or edit the.envfile, and then restart all modules. - Java command: set the key with the
-Djasypt.encryptor.passwordproperty. - Docker, Kubernetes and Cloud Run: set
ENCRYPTION_WORDin the stack.envfile or the secret. - AWS Marketplace: update the
-Djasypt.encryptor.passwordproperty in theinfo_connect_hub-3.0.0.conffile in/opt/info-connect-hub/executable-jars.
Restart the Hub after you change the key.
Changing the default user ID and password
The Admin API and the Functional API each have their own user. You can change the credentials with the credential update script or manually.
Warning
The script and the properties set the credentials of the users that the Hub creates on its first start. The users are stored in the Hub metadata database, so after the first start a new password in the properties has no effect, and a new user ID creates an additional user while the old one stays active. To change the password of an existing user, call PUT /admin/users/{username}/password. To remove an old user, call DELETE /admin/users/{username}. See User groups and permissions.
Recommended approach
Use the credential update script in <hub-home>/start-scripts:
- Windows: double-click
change-default-creds.bat. -
Linux and macOS: open a terminal, go to the
<hub-home>/start-scriptsdirectory and run:
The script asks you to enter and confirm the new credentials and updates all Hub components.
Manual configuration
- Open
<project-folder>/config/application.properties. - Find the Basic Authentication section. The user properties apply to the Functional API and the admin properties apply to the Admin API.
- To change the user IDs, update
userUsernameandadminUsername. - To change the passwords, update
userPasswordandadminPassword. -
Optional: store the passwords in encrypted form.
-
Encrypt the password with the Admin encryption API:
-
Wrap the encrypted value in
ENC(...). - Paste the result into the
userPasswordoradminPasswordproperty.
-
-
Restart the Hub.
Authentication and access control
- OAuth. Use the authorization server to authenticate with an OAuth provider instead of Basic authentication. See OAuth configuration.
- User groups and permissions. Assign users to groups and give each group access to specific Program Calls, Data APIs and RPA Calls. The same permissions apply to the generated MCP tools. See User groups and permissions.
- TLS to IBM i. Use an encrypted connection between the Hub and IBM i. See IBM i TLS connection configuration.
Availability and scaling
- Several nodes. Run several Hub nodes behind a load balancer. See Cluster configuration.
- Containers. Deploy with Docker, Kubernetes, Azure AKS, Google GKE or Google Cloud Run. See Deploying InfoConnect Hub with Docker, Kubernetes and Cloud Run.
- Data Flow throughput. Tune the batch size of a Data Flow. The Hub writes each batch to a JDBC target in one transaction, and one Data Flow can replicate many tables from the same data queue. See How the JDBC sink writes data and Replicating many tables with one Data Flow.
- One destination per table. Use dynamic data publishing to route each source table to its own queue, topic or file folder. See Dynamic data publishing.
Monitoring and recovery
- Logs. Each module writes an output log and an error log to the
logfolder. See Log files. - Health checks. The Hub and the MCP server provide
/actuator/health. The Hub health check reports only that the process is up. It does not check the metadata database or the connection to IBM i. On the MCP server the endpoint requires the MCP credentials unlessmcp.auth.type=none. See Health check. - Data Flow states. The state of a Data Flow (
ACTIVE,INACTIVE,FAILED,INVALID) shows whether it runs. See Data Flow states and errors. - Dead letter queue. A Data Flow has DLQ Settings, where you enable a dead letter queue for failed messages. See Data Flows.
- Log aggregation. The log files are plain files in a known folder, so your standard log shipper can forward them to platforms such as Splunk or the Elastic stack.