This is the documentation for Cloudera Manager 5.1.0.
Documentation for other versions is available at Cloudera Documentation.

Events

An event is a record that something of interest has occurred – a service's health has changed state, a log message (of the appropriate severity) has been logged, and so on. Many events are enabled and configured by default.

From the Events page you can filter for events for services or role instances, hosts, users, commands, and much more. You can also search against the content information returned by the event.

The Event Server aggregates relevant events and makes them available for alerting and for searching. This way, you have a view into the history of all relevant events that occur cluster-wide.

Cloudera Manager supports the following categories of events:
Category Description
ACTIVITY_EVENT Generated by the Activity Monitor; specifically, for jobs that fail, or that run slowly (as determined by comparison with duration limits). In order to monitor your workload for slow-running jobs, you must specify Activity Duration Rules.
AUDIT_EVENT Generated by actions performed
  • In Cloudera Manager, such as creating, configuring, starting, stopping, and deleting services or roles
  • By services that are being audited by Cloudera Navigator.
HBASE Generated by HBase with the exception of log messages, which have the LOG_MESSAGE category.
HEALTH_CHECK Indicate that certain health test activities have occurred, or that health test results have met specific conditions (thresholds).

Thresholds for various health tests can be set under the Configuration tabs for HBase, HDFS, Impala, and MapReduce service instances, at both the service and role level. See Configuring Health Monitoring for more information.

LOG_MESSAGE Generated for certain types of log messages from HDFS, MapReduce, and HBase services and roles. Log events are created when a log entry matches a set of rules for identifying messages of interest. The default set of rules is based on Cloudera experience supporting Hadoop clusters. You can configure additional log event rules if necessary.
SYSTEM Generated by system events such as parcel availability.

Viewing Events

The Events page lets you display events and alerts that have occurred within a time range you select anywhere in your clusters. From the Events page you can filter for events for services or role instances, hosts, users, commands, and much more. You can also search against the content information returned by the event.

To view events, click the Diagnostics tab on the top navigation bar, then select Events.

Events List

Event entries are ordered (within the time range you've selected) with the most recent at the top. If the event generated an alert, that is indicated by a red alert icon () in the entry.

This page supports infinite scrolling: you can scroll to the end of the displayed results and the page will fetch more results and add them to the end of the list automatically.

To display event details, click Expand at the right side of the event entry.

Clicking the View link at the far right of the entry has different results depending on the category of the entry:
  • ACTIVITY_EVENT - Displays the activity Details page.
  • AUDIT_EVENT - If the event was a restart, displays the service's Commands page. If the event was a configuration change, the Revision Details dialog displays.
  • HBASE - Displays a health report or log details.
  • HEALTH_CHECK - Displays the status page of the role instance.
  • LOG_MESSAGE - Displays the event's log entry. You can also click Expand to display details of the entry, then click the URL link. When you perform one of these actions the time range in the Time Line is shifted to the time the event occurred.
  • SYSTEM - Displays the Parcels page.

Filtering Events

You filter events by selecting a time range and adding filters.

You can use the Time Range Selector or a duration link () to set the time range. (See Time Line for details). The time it takes to perform a search will typically increase for a longer time range, as the number of events to be searched will be larger.

Adding a Filter

To add a filter, do one of the following:
  • Click the icon that displays next to a property when you hover in one of the event entries. A filter containing the property, operator, and its value is added to the list of filters at the left and Cloudera Manager redisplays all events that match the filter.
  • Click the Add a filter link. A filter control is added to the list of filters.
    1. Choose a property in the drop-down list. You can search by properties such as Username, Service, Command, or Role. The properties vary depending on the service or role.
    2. If the property allows it, choose an operator in the operator drop-down list.
    3. Type a property value in the value text field. For some properties you can include multiple values in the value field. For example, you can create a filter like Category = HEALTH_CHECK LOG_MESSAGE. To drop individual values, click the to the right of the value. For properties where the list of values is finite and known, you can start typing and then select from a drop-down list of potential matches.
    4. Click Search. The log displays all events that match the filter criteria.
    5. Click to add more filters and repeat steps 1 through 4.
  Note: You can filter on a string by adding a filter, selecting the property CONTENT, operator =, and typing the string to search for in the value field.

Removing a Filter

  1. Click the at the right of the filter. The filter is removed.
  2. Click Search. The log displays all events that match the filter criteria.

Re-running a Search

To re-run a recently performed search, click to the right of the Search button and select a search.