In this demo, Henry Robinson, a software engineer at Cloudera, discusses the Log Management, Event Management and Alerting features in Cloudera Manager that help make sense out of all the discrete events that take place across the Hadoop cluster. He demonstrates how to search the logs valuable information, note important events that pertain to system health and create alerts to warn you when things go wrong.
Every process in a Hadoop cluster regularly writes to a log file, which captures valuable data but also creates volumes of information that is difficult to manually sort. Cloudera Manager’s comprehensive log management feature contextualizes all system logs from across the Hadoop cluster and allows the operator to search and filter by service, role, host, keyword and severity. The application also proactively scans the log files for irregularities and warns you before the Hadoop cluster is impacted.
With event management, Cloudera Manager proactively reports on important events in the Hadoop cluster such as a change in service health or metrics, log messages with a certain severity or keyword, or abnormal job performance. It creates and aggregates these relevant Hadoop events, and makes them available for searching and alerting.
Cloudera Manager aggregates events for easy filtering and viewing and makes them available as email alerts through the Alerting feature. This powerful feature warns you before harm is done to the cluster.