You are viewing the documentation for Blueriq 16. Documentation for other versions is available in our documentation directory.

Since Blueriq 14.6, the Timeline DAO, which used to be included in the Trace SQL Store Component, has been split to a separate component. See Timeline SQL Store component for details.



Table of contents 

Description

This Trace SQL Store Component contains a Trace DAO implementation that can be used by the trace engine to store trace messages in a database.

Blueriq 14.8 introduced the possibility to consume trace events that are pushed on a queue by the Trace Event Publisher Component and then store them in a database in an asynchronous way. See consume trace events.

Requirements

This component persists data in a database, that must be provided. See platform support for the supported databases.

Optional: the consumer functionality needs to be able to access a queue. Currently only RabbitMQ is supported. For setting this up see Configuration of RabbitMQ for Customerdata service and Trace Event Publisher.

Installation

In order to use this component, the trace-sql-store profile must be active. More information on how to configure the application using Spring Profiles can be found here: Component configuration with Spring Profiles.

This component will only write trace message to the database and not timeline messages. See Timeline SQL Store component for writing timeline messages to a separate database.

Add the artifact with groupId "com.blueriq" and artifactId "blueriq-component-trace-sql-store" as a dependency to your Blueriq runtime POM file if you want to depend on it.

Configuration

The datasource name is trace-sql-store.

Database

Scripts to create the required database content are provided for the supported databases.

For customers that are upgrading, if the database scheme was already created in a previous version, check to see if there are database upgrade scripts provided for any of the intermediate versions in the Upgrade Instructions.

Configuration

Use the following steps to configure the datasource:

  1. Install the driver for your database. See Configuring JDBC database drivers for information on how to do this.
  2. Provide the connection details for the datasource. There are two options for this: using JDBC or JNDI.
    1. Using JDBC datasource : this is configured in the  application-externaldatasources.properties file. When configuring external datasources, the externaldatasources profile should be enabled.

      application-externaldatasources.properties
      blueriq.datasource.<datasource-name>.url=<JDBC url>
      blueriq.datasource.<datasource-name>.username=<username>
      blueriq.datasource.<datasource-name>.password=<password>
      blueriq.datasource.<datasource-name>.driverClassName=<driver-class>


    2. Using JNDI datasource : this can be configured in the  application-jndidatasources.properties file. When configuring JNDI datasources the jndidatasources profile should be enabled.

      blueriq.datasource.<datasource-name>.jndiName=<JNDI url>


  3. Provide the appropriate Hibernate settings in the same properties file as step 2.
    blueriq.hibernate.<datasource-name>.dialect=<database-dialect>
    blueriq.hibernate.<datasource-name>.use_nationalized_character_data=true
    blueriq.hibernate.<datasource-name>.hbm2ddl.auto=validate                 # validate is the only supported value here


Example configuration

Below are configuration examples which contain a JDBC url, driver-class and dialect for each supported database.

Oracle
blueriq.datasource.<datasource-name>.url=jdbc:oracle:thin:@<host>:<port>/<servicename>
blueriq.datasource.<datasource-name>.username=<username>
blueriq.datasource.<datasource-name>.password=<password>
blueriq.datasource.<datasource-name>.driverClassName=oracle.jdbc.driver.OracleDriver

blueriq.hibernate.<datasource-name>.dialect=org.hibernate.dialect.OracleDialect
blueriq.hibernate.<datasource-name>.hibernate.use_nationalized_character_data=true
blueriq.hibernate.<datasource-name>.hbm2ddl.auto=validate


SQL Server
blueriq.datasource.<datasource-name>.url=jdbc:sqlserver://<host>:<port>;databaseName=<database>;trustServerCertificate=true
blueriq.datasource.<datasource-name>.username=<username>
blueriq.datasource.<datasource-name>.password=<password>
blueriq.datasource.<datasource-name>.driverClassName=com.microsoft.sqlserver.jdbc.SQLServerDriver

blueriq.hibernate.<datasource-name>.dialect=org.hibernate.dialect.SQLServer2012Dialect
blueriq.hibernate.<datasource-name>.hibernate.use_nationalized_character_data=true
blueriq.hibernate.<datasource-name>.hbm2ddl.auto=validate


PostgreSQL
blueriq.datasource.<datasource-name>.url=jdbc:postgresql://<host>:<port>/<database>
blueriq.datasource.<datasource-name>.username=<username>
blueriq.datasource.<datasource-name>.password=<password>
blueriq.datasource.<datasource-name>.driverClassName=org.postgresql.Driver

blueriq.hibernate.<datasource-name>.dialect=org.hibernate.dialect.PostgreSQLDialect
blueriq.hibernate.<datasource-name>.hibernate.use_nationalized_character_data=true
blueriq.hibernate.<datasource-name>.hbm2ddl.auto=validate



Consume trace events

The Trace SQL Store component can store trace messages directly in the database, but it is also possible to decouple the storage of trace messages using the Trace Event Publisher Component to publish trace events to a queue and then consume them using this component.

To be able to consume event messages, the event channel needs to be configured in the application.properties. This file can be found, or otherwise should be placed in the

spring.config.additional-location

The Trace SQL Store either stores trace messages in the database, or consumes trace events from the queue. When the configuration below is enabled, only trace events are consumed, no trace messages are saved to the database directly. So if you enable this configuration, be sure to also enable the Trace Event Publisher Component, otherwise no trace information is saved anywhere.

The queue configuration can be configured using a default queue (using blueriq.default.rabbitmq.* in the application.properties). When it is needed, it can be overruled using the configuration below.

Also the trace events exchange has a default, and can be overruled by configuration when needed.

# RabbitMQ configuration
blueriq.trace.event.listener.amqp.rabbitmq.host=localhost
blueriq.trace.event.listener.amqp.rabbitmq.port=5672
blueriq.trace.event.listener.amqp.rabbitmq.virtualHost=/
blueriq.trace.event.listener.amqp.rabbitmq.username=guest
blueriq.trace.event.listener.amqp.rabbitmq.password=guest
blueriq.trace.event.listener.amqp.rabbitmq.queueNames=dcmTraceEvents

blueriq.trace.event.listener.amqp.concurrency.max-concurrent-consumers=1
blueriq.trace.event.listener.amqp.concurrency.concurrent-consumers=1

# Enable consumer functionality
blueriq.trace.event.listener.amqp.enabled=true

In certain circumstances it is possible that an error happens while consuming the trace event. Because of the workings of RabbitMQ, this will result in an endless loop where it tries to re-send the trace event until it is correctly consumed. To prevent this, a republish mechanism is added with Blueriq 15.7. This mechanism sends trace events to a dead-letter-queue when an error occurs. This will prevent the creation of the endless loop. For information see Configuring RabbitMQ

For more information on concurrent consumers see Configuring RabbitMQ.

Enable TLS connection

To enable a secure (TLS) AMQP connection to the queue, set the property "blueriq.trace.event.listener.amqp.rabbitmq.ssl.enabled" to "true".
The default for AQMP port with TLS on RabbitMQ is '5671', it needs to be explicitly enabled though, see https://www.rabbitmq.com/ssl.html for more information.

Performance

The performance impact of the Trace SQL Store is minimal when used normally: to write trace messages to the database.

It is also possible to use the Trace SQL Store to query the trace engine database, like the timeline container does. The Trace SQL Store offers good performance as long as the result set is limited. For large scale analysis, when summaries or aggregates are desired, the Trace SQL Store is not recommended because it is primarily designed to return a list of trace entries. For these scenarios it is recommended to use SQL (or HQL) to query the database.