Incoming connections enable IBM Cloud Pak for Watson AIOps to collect log, metric, and event data from various sources.
IBM Cloud Pak for Watson AIOps uses this collected data to establish baseline models during AI model training. These models are then used to compare against live incoming data from these connections to report anomalous activity to you in near real time by way of your outgoing connections. In live data mode, the connector might need a delay to offset the query time windows to provide time buffer to prevent retrieving partial real-time data. For more information, see Delay configuration in data connections.
In the Connector UI, there is an environment variable that allows you to optionally configure how often your connection statuses are refreshed in the UI. If you do not modify this value, the UI will use the default of 300 seconds (5 minutes). To customize this value, see Optional caching configuration for Connector UI.
The recommended value for source parallelism should be equal to the number of days. For example, if you are pulling historical data for training for 7 days, the source parallelism should be around 7 or higher, so that you can pull data for multiple days simultaneously.
Similarly for the base parallelism, it is recommended to use a higher value than 1 so that you can process data in parallel. In a small environment the available flinks slots is 16 and for a large environment, the maximum available slots is 32.
Unlike outgoing connections, you can set up multiple sources of incoming connections. For example, if you have a single source of Falcon LogScale data for logs, and PagerDuty configured for events, you can create connections for both of those. You can also use the custom and Kafka integrator for greater flexibility. The custom connection type collects data from a single specific data source (like any other connection type). If you want to collect log data from different systems, and then route that information through a forwarding agent, such as an Apache Kafka topic (like Sysdig or FluentBit), you can use the Kafka connection type.
Note: The ServiceNow connection type has optional incoming and outgoing facets.
For more information about incoming connections, see the following topics:
- Administrator process connections
- Ansible connections
- Appdynamics connections
- AWS CloudWatch connections
- Custom connections
- Custom connectors
- Dynatrace connections
- ELK connections
- Falcon LogScale connections
- Generic webhook connection
- IBM Tivoli Netcool/Impact connection
- Instana connections
- Kafka connections
- LogDNA connections
- Netcool connections
- New Relic connections
- Observer connections
- PagerDuty connections
- Splunk connections
- SSH connections
- Zabbix connections