Known Issues
This section lists known issues and workarounds in Cloudera Observability on premises 3.7.1.
- LDAP authentication fails when the LDAP server certificate has an empty Subject field
- The Cloudera Observability Console Server cannot authenticate users when the TLS certificate of the LDAP server contains an empty Subject field, even if the certificate contains valid DNS subject alternative names. Authentication fails during the initial LDAP bind, and the Console Server log displays the following error: Error [ERR_TLS_CERT_ALTNAME_INVALID]: Hostname/IP does not match certificate's altnames: Cert is empty. This issue affects Cloudera Observability on premises 3.7.1 and all hotfix builds lower than 3.7.1-h4.
- SSL handshake failures
- In Cloudera Data Services on premises environments, telemetry agent pods, such as database producer in Cloudera Data Warehouse or dbus-wxml-client in Cloudera Data Engineering can encounter an SSLHandshakeException when communicating with the DBus API. The pod logs display a PKIX path building failed error.
- Impala does not support super user configuration for the observability user for Apache Ranger-enabled cluster
- The observability user requires full privileges on the Observability cluster. Required services such as Kafka, HDFS, HBase, and Hive support super user setup by specifying the ranger.plugin.[service].super.users property to observability. However, this super user setup is not supported for Impala.
- Exporting of Impala queries fail for Telemetry Publisher with Cloudera Manager 7.11.3
- Telemetry Publisher for Impala queries does not work with Cloudera Manager 7.11.3
- Auto Action trigger for Impala Engine
- Impala Auto Action triggers do not work for the Kerberos-enabled Private Cloud base cluster running on Cloudera Manager 7.9.5 and 7.11.3.
- Telemetry publisher test altus connection fails for Cloudera Manager 7.11.3 hotfix (CHF6, 7, and 8) versions
- Test connection fails with the following error:
Exception in thread "main" java.lang.NoSuchMethodError: 'com.google.common.collect.ImmutableSet com.google.common.collect.ImmutableSet.copyOf(java.util.Collection)' at com.cloudera.cdp.http.HttpCodesRetryChecker.<init>(HttpCodesRetryChecker.java:57) at com.cloudera.cdp.client.CdpClientConfigurationBuilder.<init>(CdpClientConfigurationBuilder.java:53) at com.cloudera.cdp.client.CdpClientConfigurationBuilder.defaultBuilder(CdpClientConfigurationBuilder.java:400) at com.cloudera.cdx.client.TestDatabusConnection.main(TestDatabusConnection.java:55)This issue only affects the test connection method.
- Limitations of Cloudera Observability Spark processing
- Spark event log processing is skipped under the following conditions:
- The event log size exceeds 20 GB.
- The number of Spark jobs in the Spark application exceeds 500.
- The number of SQL executions in the Spark application exceeds 500.
- When spark.history.fs.eventLog.rolling.maxFilesToRetain is set, Spark keeps only recent event log files as configured inspark.history.fs.eventLog.rolling.maxFilesToRetain and older event log files are compacted (in a file with .compact extension). Spark event log processing of such Spark applications are skipped by Cloudera Observability.
