Adaptive Data Agents
The Adaptive data agent is a component of Adaptive Integration that runs on a Adaptive Insights customer on-premises server. The data agent extracts information from a JDBC-compliant database or a custom scripted data source. Data agents can also export data from Adaptive Insights. The data agent requires a computer running Windows deployed behind a customer firewall, and runs as a hosted service. The data agent manages access to on-premises applications as well as connections to Adaptive Integration on the cloud.
The Adaptive Data Agent Service (the Agent Service) manages the data agents. This service runs as a Windows Service and launches a Java application installed behind the customer’s firewall. The Agent Service long polls the Adaptive Gateway. The Adaptive Gateway server brokers communication through the firewall to on-premises instances of the Agent Service. This gateway receives requests from Adaptive Insights servers on behalf of all data agents hosted by that instance of the service. Each installed instance of the Agent Service can host multiple data agents.
The installation includes the Adaptive Data Agent Service Manager, a desktop utility the customer uses to intially setup and configure the Agent Service. The installation optionally includes Pentaho Kettle and Spoon, the Kettle ETL Designer. If the optional Pentaho Kettle components get used, the Agent Service manages a pool of Kettle ETL execution instances (job runners). Each execution instance is a Java virtual machine running in its own process. The Agent Service delegates any Kettle ETL scripts that launch to a free job runner. Adaptive Insights provides plugins to enable design-time connectivity between the Spoon environment and Adaptive Integration. These plugins allow for authenticating into a cloud instance of Adaptive Insights, publishing scripts to the cloud, and testing the script-based integration from within the Kettle UI. Note that the optional Pentaho components are needed only if there is a requirement to use a Pentaho Kettle script to extract data. These components are not needed if the data agent extracts data out of JDBC-compliant databases using Adaptive Integration’s JDBC data sources.
The on-premises data agents get managed and controlled from the cloud using the agent UI. The agent UI is part of Adaptive Integration’s data designer, which is accessed through a web browser. When an agent is created in the data designer, the customer can download and install the latest version of the data agent, along with the optional Pentaho Kettle components. The agent UI shows the current status and version details of the agents and allows the customer to provision/unprovision/suspend/resume data agents. The customer can also upgrade to a newer version of the data agent through the agent UI. See Adaptive Integration.
The following technical specifications are recommended for the machine the data agent will reside on. Exact specifications will depend on data volumes and should be reviewed with an Adaptive Insights Professional Services consultant.
- Windows 2019 Server
- Windows 2016 Server
- Windows 2012 Server (R2)
- Windows 2008 Server (R2 is recommended)
- Windows 10
- Windows 8.1
- Windows 7
Data Agent Support for Pentaho Kettle and Java version:
|Data Agent Version||Java Version||Pentaho Data Integration (Kettle) Version|
|2019.3 Data Agent v8||8, 11, OpenJDK||8.2|
|2017.1 Data Agent v60 and v7x||8||6.1|
|Legacy Data Agent earlier than v60||7||4.4|
2019.3 Data Agent v8 Support by OS, JDK, and Kettle Version
|OS||JDK||Pentaho Data Integration (Kettle) Version|
|Windows Server 2016||Open JDK 8||No Kettle|
|Windows Server 2016||Open JDK 8||Kettle 8.2|
|Windows Server 2016||Open JDK 8||Kettle 7.1|
|Windows Server 2016||Open JDK 8||Kettle 6.1|
|Windows Server 2016||Open JDK 11||No Kettle|
|Windows Server 2019||Open JDK 8||No Kettle|
|Windows Server 2019||Open JDK 11||No Kettle|
|Windows Server 2016||Oracle JRE 8||No Kettle|
|Windows Server 2016||Oracle JDK 8||Kettle 8.2|
|Windows Server 2016||Oracle JDK 8||Kettle 7.1|
|Windows Server 2016||Oracle JDK 8||Kettle 6.1|
|Windows Server 2016||Oracle Java SE 11||No Kettle|
|Windows Server 2019||Oracle Java SE 8||No Kettle|
|Windows Server 2019||Oracle Java SE 11||No Kettle|
Kettle Spoon starts with a parameter setting of Max Heap Size 2048MB if you don't specify your own. When Spoon launches with a 32-bit JDK, this parameter prevents successful start-up. For 32-bit JDKs set the Max Heap Size using:
60 GB of disk space is recommended. This is used for temporary files and disk caching and actual size needed will vary based on the needs of the customer’s implementation.
A minimum of 8GB is recommended. If a virtual machine is used to host the data agent, 12GB is recommended.
The data agent uses HTTPS to communicate with the Adaptive Insights servers. Port 443 (default port for HTTPS) is used by the Data Agent Service Manager to communicate with the Adaptive Insights servers and other web-based sources.
One machine running the agent service cannot serve tenants in multiple geographies.
Traffic balancing and data sovereignty/data residency management requires whitelisting