Monitoring Opportunity Detect deployments on Streams

Opportunity Detect provides a monitoring tool for deployments running on Streams.

The monitoring tool runs as a JConsole service.

The data in the monitoring tool is supplied by the Streams Remote Control Service. Only running deployments and services are shown.

You can view the following information in real time.

  • Health of the Streams jobs
  • Metrics for individual Opportunity Detect deployments within Streams
  • Aggregated metrics for servers, server groups and Opportunity Detect deployments

Notifications

Two types of notification are available.

  • A Streams service or job (deployment) failure can trigger a notification to users who have Opportunity Detect roles. Users who have Opportunity Detect roles and who subscribe to Marketing Platform system alerts receive these notifications, which include details, in their HCL® Marketing Software inbox.

    If email notifications are configured, subscribing Opportunity Detect users also receive an email for each failure.

  • Users of the monitoring tool can also subscribe to notifications that appear in the tool, using the JConsole notifications feature.

Problem diagnosis

The monitoring tool helps you to diagnose problems by providing information on the following types of failure.

  • Streams Remote Control Service (RCS)

    The Streams RCS obtains metrics. There is a single instance of the RCS across all server groups. After an RCS failure, the RCS is unable to obtain metrics.

  • Streams Real Time Web Connector (RTC)

    The RTC is the incoming gateway for real-time traffic. There is a single instance of the RTC for each server. After an RTC failure, the load balancer routes incoming traffic to other RTC instances that are running.

  • Streams Real Time Proxy (RTP)

    There is a single instance of the Streams RTP for each server. The RTP routs incoming messages to engines. An RTP failure causes loss of incoming messages.

  • Database failure

    When there is a database failure, engines fail to update State History. This causes engines to fail when they exceed their in-memory cache limit or the number of allowed connection retries.

  • Engine failure

    After an engine failure, transactions targeted to the failed engine are no longer processed.

  • Streams input/output jobs

    Input/output jobs are used in the operation of the Opportunity Detect batch feeder and listener. After a Streams failure, parts of the input/output chain fail.