Events are records of certain occurrences in the system and are very useful to track the different operations running on the cluster. Several events are configured by default when Cloudera Manager is installed.
All events in the cluster are managed by the Event Server component of Cloudera Manager. To view the events, navigate to Diagnostics | Events from the Cloudera Manager toolbar. The following is the screenshot of the Events page:
The preceding screenshot shows one event for the HDFS service that is part of the node1.hcuster host and the datanode (node1) role. The screenshot also shows the details of the alert that is displayed when you click on the Expand link. To close the details, you can use the Collapse link.
Search filters can also be applied to perform more advanced searches. Click on the Add a filter hyperlink to add filters as shown in the following screenshot:
To open up the datanode instance, click on the View button. The following screenshot shows the datanode (node1) instance that corresponds to the event:
To view the logfile for this datanode, that is, datanode (node1), click on the Log File link to open the Logs search screen as shown in the following screenshot:
Alerts are messages that are configured to be triggered by an event. Alerts are configured to inform the occurrence of events that need special attention.
The following are the steps to configure an event. For this example, we will configure an event to monitor HDFS free space:
- Navigate to the Home page from the Cloudera Manager toolbar.
- Select the HDFS service to navigate to the HDFS service details page.
- Navigate to Configuration | View and Edit and select the Monitoring section.
- Click on Service-Wide to display all the configuration parameters for the HDFS service as shown in the following screenshot:
- The default value configured for the HDFS Free Space Monitoring Thresholds property are 20.0% and 10.0% for the Warning and Critical events, respectively. This means that a Warning event will be triggered when the free space on HDFS reaches 20 percent and a Critical event will be triggered when the HDFS free space reaches 10 percent.
For this illustration, let's update the values 25% and 20% for the Warning and Critical events respectively, as shown in the following screenshot:
- Once done, save the changes by clicking on the Save Changes button.
- Now, whenever the free space on HFDS goes below 25 percent, a Warning event will be generated. Similarly, if the free space on HDFS goes below 20 percent, a Critical event will be generated.
The following screenshot shows the event generated when the free space on HDFS is below the set critical threshold of 20 percent:
You will observe that the event is labeled as Alert. This is because the HDFS service is configured to alert us if the service's status becomes bad.
- To see more details for this alert, click on the Expand link. You should see additional details as shown in the following screenshot:
- To see the list of all alerts configured, navigate to the Administration menu of the Cloudera Manager toolbar and select Alerts. The Alerts page is displayed as shown in the following screenshot:
The following alert types can be configured from Cloudera Manager:
- Health: The properties configured within this section are to do with the health of the individual services running within the cluster.
- Log: The properties configured within this section are to do with alerts to be generated for certain types of log messages found in the logs of the individual services.
- Activity: The properties configured within this section are to do with alerts to be generated by the activity monitor that monitors jobs that fail or perform slowly.
Administrators would want to be alerted of certain events by e-mail rather than having to check the Events page for alerts and events. The following are the steps to configure the alert delivery by an e-mail:
- Navigate to the Home screen from the Cloudera Manager toolbar.
- Click on the Cloudera Management Service tab to open up the Cloudera Management Service details page and navigate to Configuration | View and Edit as shown in the following screenshot:
- Select the Alert Publisher Default Group section as shown in the following screenshot:
- Update the Alerts: Mail Server Hostname property to the IP address or the hostname of your SMTP server.
- Update the Alerts: Mail Message Recipients property to the e-mail ID or a comma-separated list of e-mail IDs to whom the alerts need to be sent, as shown in the following screenshot:
- Click on Save Changes to save the configuration changes.
- Restart the Cloudera Manager service for the changes to take effect.
- We are done with the configurations and will now start receiving alerts via e-mail. A sample alert e-mail is shown in the following screenshot:
Tip
In this chapter, we only covered the basics of monitoring, events, and alerts. Cloudera documentation provides an in-depth description of the various options available for monitoring the entire cluster and can be found at http://www.cloudera.com/content/cloudera-content/cloudera-docs/CM5/latest/Cloudera-Manager-Diagnostics-Guide/Cloudera-Manager-Diagnostics-Guide.html.