Understanding Validations in Cloudera Observability
Learn about Cloudera Support’s Validations predictive alerting tool whose alerts are displayed in the Cloudera Observability web UI. The Validations feature identifies problems within your environment arising from your support bundles and automatically displays an alert when an issue appears and/or when conditions are not met, including details about the problem and recommended solutions that enable you to take corrective action before a more serious issue arises.
Cloudera Observability provides access to the Cloudera Support's Validations predictive alerts through its Validations feature. The Validations predictive alerting tool uses predictive checks, known as validators, that are automatically run against every diagnostic bundle that is received from a supported CDH or CDP Private Cloud Base environment.
The validators analyze and identify known problem signatures contained in the diagnostic data of your bundles, such as the state and configuration settings of your cluster. There are over 400 validator problem signatures relating to misconfigurations, security vulnerabilities, performance degradation, and deviations from Cloudera’s known best practices, as well as many other types of commonly experienced issues that can affect the stability, performance, security, and health of your environment.
The Cluster validation signatures also ensure that Cloudera Data Hub (CDH) clusters are optimized for a successful upgrade to Cloudera Data Platform (CDP). Where, configuration best practices that are required for a smooth upgrade are reviewed and an alert is raised accordingly on each cluster asset if it’s out of alignment.
- Diagnostic bundle, which is created by Cloudera Manager and contains all the configurations, logs, and details about your cluster and its services.
- Application bundle, which is created from your applications, such as the Spark application, a specific Hive query, or a specific item, such as a workload job execution.
Understanding the severity values
- Critical, which indicates a serious problem that must be resolved immediately.
- Error, which indicates incorrect settings and configurations that require attention.
- Warn, which indicates a potential problem that eventually must be resolved but does not have to be completed at this time.
- Info, which is displayed for informational purposes only, such as performance improvements. For example, to handle all services, configure your log and query redactions in Cloudera Manager rather than in HDFS.
- Curiosity, which flags unusual deployments and settings. For example, Cloudera is curious as to why this type of file system is used.
- Pass, which indicates a normal result and within the acceptable range.
- Insufficient_information, which indicates that there is not enough information at this time.
Based on the Validator’s threshold severity values, the most serious alert issues are displayed in Cloudera Observability, including details about the identified error, its cause, severity level, affected hosts, the corrective actions you should consider to resolve the problem, and links to applicable documentation.
Considerations and limitations
- At this time your validations critical alerts are only available for CDP Data Hub, CDP Private Cloud Base, and Classic Cluster environments.
- The Validations feature requires that a support bundle has been sent to Cloudera Support
through one of the following methods:
- Directly, as a scheduled diagnostic delivery of diagnostic data from CDP Private Cloud Base with Cloudera Manager.
- Attachment, as a support bundle attachment in a Support case.
- Manually, through a phone home python script that is available on Cloudera.com.
For more information, click the Related Information links below.
- New and updated validator signatures are continuously added to the Validations feature by Cloudera Support. To have the latest signatures run on your bundles, you must regularly send your support bundles to Cloudera.