×

Logging is provided as an installable component, with a distinct release cycle from the core OpenShift Container Platform. The Red Hat OpenShift Container Platform Life Cycle Policy outlines release compatibility.

The stable channel only provides updates to the most recent release of logging. To continue receiving updates for prior releases, you must change your subscription channel to stable-x.y, where x.y represents the major and minor version of logging you have installed. For example, stable-5.7.

Logging 5.7.15

This release includes OpenShift Logging Bug Fix 5.7.15.

Bug fixes

  • Before this update, there was a delay in restarting Ingesters when configuring LokiStack, because the Loki Operator sets the write-ahead log replay_memory_ceiling to zero bytes for the 1x.demo size. With this update, the minimum value used for the replay_memory_ceiling has been increased to avoid delays. (LOG-5616)

Logging 5.7.14

This release includes OpenShift Logging Bug Fix 5.7.14.

Bug fixes

  • Before this update, an issue in the metrics collection code of the Logging Operator caused it to report stale telemetry metrics. With this update, the Logging Operator does not report stale telemetry metrics. (LOG-5472)

Logging 5.7.13

This release includes OpenShift Logging Bug Fix 5.7.13.

Enhancements

  • Before this update, the Loki Operator configured Loki to use path-based style access for the Amazon Simple Storage Service (S3), which has been deprecated. With this update, the Loki Operator defaults to virtual-host style without users needing to change their configuration. (LOG-5403)

  • Before this update, the Loki Operator did not validate the Amazon Simple Storage Service (S3) endpoint used in the storage secret. With this update, the validation process ensures the S3 endpoint is a valid S3 URL, and the LokiStack status updates to indicate any invalid URLs. (LOG-5393)

Bug fixes

  • Before this update, the Elastisearch Operator ServiceMonitor in the openshift-operators-redhat namespace used static token and certificate authority (CA) files for authentication, causing errors in the Prometheus Operator in the User Workload Monitoring specification on the ServiceMonitor configuration. With this update, the Elastisearch Operator ServiceMonitor in the openshift-operators-redhat namespace now references a service account token secret by a LocalReference object. This approach allows the User Workload Monitoring specifications in the Prometheus Operator to handle the Elastisearch Operator ServiceMonitor successfully. This enables Prometheus to scrape the Elastisearch Operator metrics. (LOG-5243)

  • Before this update, the Loki Operator did not validate the Amazon Simple Storage Service (S3) endpoint URL format used in the storage secret. With this update, the S3 endpoint URL goes through a validation step that reflects on the status of the LokiStack. (LOG-5399)

Logging 5.7.12

This release includes OpenShift Logging Bug Fix 5.7.12.

Bug fixes

  • Before this update, the Loki Operator checked if the pods were running to decide if the LokiStack was ready. With this update, it also checks if the pods are ready, so that the readiness of the LokiStack reflects the state of its components. (LOG-5172)

  • Before this update, the Red Hat build pipeline didn’t use the existing build details in Loki builds and omitted information such as revision, branch, and version. With this update, the Red Hat build pipeline now adds these details to the Loki builds, fixing the issue. (LOG-5202)

  • Before this update, the configuration of the Loki Operator’s ServiceMonitor could match many Kubernetes services, resulting in the Loki Operator’s metrics being collected multiple times. With this update, the configuration of ServiceMonitor now only matches the dedicated metrics service. (LOG-5251)

  • Before this update, the build pipeline did not include linker flags for the build date, causing Loki builds to show empty strings for buildDate and goVersion. With this update, adding the missing linker flags in the build pipeline fixes the issue. (LOG-5275)

  • Before this update, the Loki Operator ServiceMonitor in the openshift-operators-redhat namespace used static token and CA files for authentication, causing errors in the Prometheus Operator in the User Workload Monitoring spec on the ServiceMonitor configuration. With this update, the Loki Operator ServiceMonitor in openshift-operators-redhat namespace now references a service account token secret by a LocalReference object. This approach allows the User Workload Monitoring spec in the Prometheus Operator to handle the Loki Operator ServiceMonitor successfully, enabling Prometheus to scrape the Loki Operator metrics. (LOG-5241)

Logging 5.7.11

This release includes Logging Bug Fix 5.7.11.

Bug fixes

  • Before this update, when configured to read a custom S3 Certificate Authority, the Loki Operator would not automatically update the configuration when the name of the ConfigMap object or the contents changed. With this update, the Loki Operator now watches for changes to the ConfigMap object and automatically updates the generated configuration. (LOG-4968)

Logging 5.7.9

Bug fixes

  • Before this fix, IPv6 addresses would not be parsed correctly after evaluating a host or multiple hosts for placeholders. With this update, IPv6 addresses are correctly parsed. (LOG-4281)

  • Before this update, the Vector failed to start on IPv4-only nodes. As a result, it failed to create a listener for its metrics endpoint with the following error: Failed to start Prometheus exporter: TCP bind failed: Address family not supported by protocol (os error 97). With this update, the Vector operates normally on IPv4-only nodes. (LOG-4589)

  • Before this update, during the process of creating index patterns, the default alias was missing from the initial index in each log output. As a result, Kibana users were unable to create index patterns by using OpenShift Elasticsearch Operator. This update adds the missing aliases to OpenShift Elasticsearch Operator, resolving the issue. Kibana users can now create index patterns that include the {app,infra,audit}-000001 indexes. (LOG-4806)

  • Before this update, the Loki Operator did not mount a custom CA bundle to the ruler pods. As a result, during the process to evaluate alerting or recording rules, object storage access failed. With this update, the Loki Operator mounts the custom CA bundle to all ruler pods. The ruler pods can download logs from object storage to evaluate alerting or recording rules. (LOG-4837)

  • Before this update, changing a LogQL query using controls such as time range or severity changed the label matcher operator as though it was defined like a regular expression. With this update, regular expression operators remain unchanged when updating the query. (LOG-4842)

  • Before this update, the Vector collector deployments relied upon the default retry and buffering behavior. As a result, the delivery pipeline backed up trying to deliver every message when the availability of an output was unstable. With this update, the Vector collector deployments limit the number of message retries and drop messages after the threshold has been exceeded. (LOG-4536)

Logging 5.7.8

Bug fixes

  • Before this update, there was a potential conflict when the same name was used for the outputRefs and inputRefs parameters in the ClusterLogForwarder custom resource (CR). As a result, the collector pods entered in a CrashLoopBackOff status. With this update, the output labels contain the OUTPUT_ prefix to ensure a distinction between output labels and pipeline names. (LOG-4383)

  • Before this update, while configuring the JSON log parser, if you did not set the structuredTypeKey or structuredTypeName parameters for the Cluster Logging Operator, no alert would display about an invalid configuration. With this update, the Cluster Logging Operator informs you about the configuration issue. (LOG-4441)

  • Before this update, if the hecToken key was missing or incorrect in the secret specified for a Splunk output, the validation failed because the Vector forwarded logs to Splunk without a token. With this update, if the hecToken key is missing or incorrect, the validation fails with the A non-empty hecToken entry is required error message. (LOG-4580)

  • Before this update, selecting a date from the Custom time range for logs caused an error in the web console. With this update, you can select a date from the time range model in the web console successfully. (LOG-4684)

Logging 5.7.7

Bug fixes

  • Before this update, FluentD normalized the logs emitted by the EventRouter differently from Vector. With this update, the Vector produces log records in a consistent format. (LOG-4178)

  • Before this update, there was an error in the query used for the FluentD Buffer Availability graph in the metrics dashboard created by the Cluster Logging Operator as it showed the minimum buffer usage. With this update, the graph shows the maximum buffer usage and is now renamed to FluentD Buffer Usage. (LOG-4555)

  • Before this update, deploying a LokiStack on IPv6-only or dual-stack OpenShift Container Platform clusters caused the LokiStack memberlist registration to fail. As a result, the distributor pods went into a crash loop. With this update, an administrator can enable IPv6 by setting the lokistack.spec.hashRing.memberlist.enableIPv6: value to true, which resolves the issue. (LOG-4569)

  • Before this update, the log collector relied on the default configuration settings for reading the container log lines. As a result, the log collector did not read the rotated files efficiently. With this update, there is an increase in the number of bytes read, which allows the log collector to efficiently process rotated files. (LOG-4575)

  • Before this update, the unused metrics in the Event Router caused the container to fail due to excessive memory usage. With this update, there is reduction in the memory usage of the Event Router by removing the unused metrics. (LOG-4686)

Logging 5.7.6

Bug fixes

  • Before this update, the collector relied on the default configuration settings for reading the container log lines. As a result, the collector did not read the rotated files efficiently. With this update, there is an increase in the number of bytes read, which allows the collector to efficiently process rotated files. (LOG-4501)

  • Before this update, when users pasted a URL with predefined filters, some filters did not reflect. With this update, the UI reflects all the filters in the URL. (LOG-4459)

  • Before this update, forwarding to Loki using custom labels generated an error when switching from Fluentd to Vector. With this update, the Vector configuration sanitizes labels in the same way as Fluentd to ensure the collector starts and correctly processes labels. (LOG-4460)

  • Before this update, the Observability Logs console search field did not accept special characters that it should escape. With this update, it is escaping special characters properly in the query. (LOG-4456)

  • Before this update, the following warning message appeared while sending logs to Splunk: Timestamp was not found. With this update, the change overrides the name of the log field used to retrieve the Timestamp and sends it to Splunk without warning. (LOG-4413)

  • Before this update, the CPU and memory usage of Vector was increasing over time. With this update, the Vector configuration now contains the expire_metrics_secs=60 setting to limit the lifetime of the metrics and cap the associated CPU usage and memory footprint. (LOG-4171)

  • Before this update, the LokiStack gateway cached authorized requests very broadly. As a result, this caused wrong authorization results. With this update, LokiStack gateway caches on a more fine-grained basis which resolves this issue. (LOG-4393)

  • Before this update, the Fluentd runtime image included builder tools which were unnecessary at runtime. With this update, the builder tools are removed, resolving the issue. (LOG-4467)

Logging 5.7.4

Bug fixes

  • Before this update, when forwarding logs to CloudWatch, a namespaceUUID value was not appended to the logGroupName field. With this update, the namespaceUUID value is included, so a logGroupName in CloudWatch appears as logGroupName: vectorcw.b443fb9e-bd4c-4b6a-b9d3-c0097f9ed286. (LOG-2701)

  • Before this update, when forwarding logs over HTTP to an off-cluster destination, the Vector collector was unable to authenticate to the cluster-wide HTTP proxy even though correct credentials were provided in the proxy URL. With this update, the Vector log collector can now authenticate to the cluster-wide HTTP proxy. (LOG-3381)

  • Before this update, the Operator would fail if the Fluentd collector was configured with Splunk as an output, due to this configuration being unsupported. With this update, configuration validation rejects unsupported outputs, resolving the issue. (LOG-4237)

  • Before this update, when the Vector collector was updated an enabled = true value in the TLS configuration for AWS Cloudwatch logs and the GCP Stackdriver caused a configuration error. With this update, enabled = true value will be removed for these outputs, resolving the issue. (LOG-4242)

  • Before this update, the Vector collector occasionally panicked with the following error message in its log: thread 'vector-worker' panicked at 'all branches are disabled and there is no else branch', src/kubernetes/reflector.rs:26:9. With this update, the error has been resolved. (LOG-4275)

  • Before this update, an issue in the Loki Operator caused the alert-manager configuration for the application tenant to disappear if the Operator was configured with additional options for that tenant. With this update, the generated Loki configuration now contains both the custom and the auto-generated configuration. (LOG-4361)

  • Before this update, when multiple roles were used to authenticate using STS with AWS Cloudwatch forwarding, a recent update caused the credentials to be non-unique. With this update, multiple combinations of STS roles and static credentials can once again be used to authenticate with AWS Cloudwatch. (LOG-4368)

  • Before this update, Loki filtered label values for active streams but did not remove duplicates, making Grafana’s Label Browser unusable. With this update, Loki filters out duplicate label values for active streams, resolving the issue. (LOG-4389)

  • Pipelines with no name field specified in the ClusterLogForwarder custom resource (CR) stopped working after upgrading to OpenShift Logging 5.7. With this update, the error has been resolved. (LOG-4120)

Logging 5.7.3

Bug fixes

  • Before this update, when viewing logs within the OpenShift Container Platform web console, cached files caused the data to not refresh. With this update the bootstrap files are not cached, resolving the issue. (LOG-4100)

  • Before this update, the Loki Operator reset errors in a way that made identifying configuration problems difficult to troubleshoot. With this update, errors persist until the configuration error is resolved. (LOG-4156)

  • Before this update, the LokiStack ruler did not restart after changes were made to the RulerConfig custom resource (CR). With this update, the Loki Operator restarts the ruler pods after the RulerConfig CR is updated. (LOG-4161)

  • Before this update, the vector collector terminated unexpectedly when input match label values contained a / character within the ClusterLogForwarder. This update resolves the issue by quoting the match label, enabling the collector to start and collect logs. (LOG-4176)

  • Before this update, the Loki Operator terminated unexpectedly when a LokiStack CR defined tenant limits, but not global limits. With this update, the Loki Operator can process LokiStack CRs without global limits, resolving the issue. (LOG-4198)

  • Before this update, Fluentd did not send logs to an Elasticsearch cluster when the private key provided was passphrase-protected. With this update, Fluentd properly handles passphrase-protected private keys when establishing a connection with Elasticsearch. (LOG-4258)

  • Before this update, clusters with more than 8,000 namespaces caused Elasticsearch to reject queries because the list of namespaces was larger than the http.max_header_size setting. With this update, the default value for header size has been increased, resolving the issue. (LOG-4277)

  • Before this update, label values containing a / character within the ClusterLogForwarder CR would cause the collector to terminate unexpectedly. With this update, slashes are replaced with underscores, resolving the issue. (LOG-4095)

  • Before this update, the Cluster Logging Operator terminated unexpectedly when set to an unmanaged state. With this update, a check to ensure that the ClusterLogging resource is in the correct Management state before initiating the reconciliation of the ClusterLogForwarder CR, resolving the issue. (LOG-4177)

  • Before this update, when viewing logs within the OpenShift Container Platform web console, selecting a time range by dragging over the histogram didn’t work on the aggregated logs view inside the pod detail. With this update, the time range can be selected by dragging on the histogram in this view. (LOG-4108)

  • Before this update, when viewing logs within the OpenShift Container Platform web console, queries longer than 30 seconds timed out. With this update, the timeout value can be configured in the configmap/logging-view-plugin. (LOG-3498)

  • Before this update, when viewing logs within the OpenShift Container Platform web console, clicking the more data available option loaded more log entries only the first time it was clicked. With this update, more entries are loaded with each click. (OU-188)

  • Before this update, when viewing logs within the OpenShift Container Platform web console, clicking the streaming option would only display the streaming logs message without showing the actual logs. With this update, both the message and the log stream are displayed correctly. (OU-166)

Logging 5.7.2

Bug fixes

  • Before this update, it was not possible to delete the openshift-logging namespace directly due to the presence of a pending finalizer. With this update, the finalizer is no longer utilized, enabling direct deletion of the namespace. (LOG-3316)

  • Before this update, the run.sh script would display an incorrect chunk_limit_size value if it was changed according to the OpenShift Container Platform documentation. However, when setting the chunk_limit_size via the environment variable $BUFFER_SIZE_LIMIT, the script would show the correct value. With this update, the run.sh script now consistently displays the correct chunk_limit_size value in both scenarios. (LOG-3330)

  • Before this update, the OpenShift Container Platform web console’s logging view plugin did not allow for custom node placement or tolerations. This update adds the ability to define node placement and tolerations for the logging view plugin. (LOG-3749)

  • Before this update, the Cluster Logging Operator encountered an Unsupported Media Type exception when trying to send logs to DataDog via the Fluentd HTTP Plugin. With this update, users can seamlessly assign the content type for log forwarding by configuring the HTTP header Content-Type. The value provided is automatically assigned to the content_type parameter within the plugin, ensuring successful log transmission. (LOG-3784)

  • Before this update, when the detectMultilineErrors field was set to true in the ClusterLogForwarder custom resource (CR), PHP multi-line errors were recorded as separate log entries, causing the stack trace to be split across multiple messages. With this update, multi-line error detection for PHP is enabled, ensuring that the entire stack trace is included in a single log message. (LOG-3878)

  • Before this update, ClusterLogForwarder pipelines containing a space in their name caused the Vector collector pods to continuously crash. With this update, all spaces, dashes (-), and dots (.) in pipeline names are replaced with underscores (_). (LOG-3945)

  • Before this update, the log_forwarder_output metric did not include the http parameter. This update adds the missing parameter to the metric. (LOG-3997)

  • Before this update, Fluentd did not identify some multi-line JavaScript client exceptions when they ended with a colon. With this update, the Fluentd buffer name is prefixed with an underscore, resolving the issue. (LOG-4019)

  • Before this update, when configuring log forwarding to write to a Kafka output topic which matched a key in the payload, logs dropped due to an error. With this update, Fluentd’s buffer name has been prefixed with an underscore, resolving the issue.(LOG-4027)

  • Before this update, the LokiStack gateway returned label values for namespaces without applying the access rights of a user. With this update, the LokiStack gateway applies permissions to label value requests, resolving the issue. (LOG-4049)

  • Before this update, the Cluster Logging Operator API required a certificate to be provided by a secret when the tls.insecureSkipVerify option was set to true. With this update, the Cluster Logging Operator API no longer requires a certificate to be provided by a secret in such cases. The following configuration has been added to the Operator’s CR:

    tls.verify_certificate = false
    tls.verify_hostname = false
  • Before this update, the LokiStack route configuration caused queries running longer than 30 seconds to timeout. With this update, the LokiStack global and per-tenant queryTimeout settings affect the route timeout settings, resolving the issue. (LOG-4052)

  • Before this update, a prior fix to remove defaulting of the collection.type resulted in the Operator no longer honoring the deprecated specs for resource, node selections, and tolerations. This update modifies the Operator behavior to always prefer the collection.logs spec over those of collection. This varies from previous behavior that allowed using both the preferred fields and deprecated fields but would ignore the deprecated fields when collection.type was populated. (LOG-4185)

  • Before this update, the Vector log collector did not generate TLS configuration for forwarding logs to multiple Kafka brokers if the broker URLs were not specified in the output. With this update, TLS configuration is generated appropriately for multiple brokers. (LOG-4163)

  • Before this update, the option to enable passphrase for log forwarding to Kafka was unavailable. This limitation presented a security risk as it could potentially expose sensitive information. With this update, users now have a seamless option to enable passphrase for log forwarding to Kafka. (LOG-3314)

  • Before this update, Vector log collector did not honor the tlsSecurityProfile settings for outgoing TLS connections. After this update, Vector handles TLS connection settings appropriately. (LOG-4011)

  • Before this update, not all available output types were included in the log_forwarder_output_info metrics. With this update, metrics contain Splunk and Google Cloud Logging data which was missing previously. (LOG-4098)

  • Before this update, when follow_inodes was set to true, the Fluentd collector could crash on file rotation. With this update, the follow_inodes setting does not crash the collector. (LOG-4151)

  • Before this update, the Fluentd collector could incorrectly close files that should be watched because of how those files were tracked. With this update, the tracking parameters have been corrected. (LOG-4149)

  • Before this update, forwarding logs with the Vector collector and naming a pipeline in the ClusterLogForwarder instance audit, application or infrastructure resulted in collector pods staying in the CrashLoopBackOff state with the following error in the collector log:

    ERROR vector::cli: Configuration error. error=redefinition of table transforms.audit for key transforms.audit

    After this update, pipeline names no longer clash with reserved input names, and pipelines can be named audit, application or infrastructure. (LOG-4218)

  • Before this update, when forwarding logs to a syslog destination with the Vector collector and setting the addLogSource flag to true, the following extra empty fields were added to the forwarded messages: namespace_name=, container_name=, and pod_name=. With this update, these fields are no longer added to journal logs. (LOG-4219)

  • Before this update, when a structuredTypeKey was not found, and a structuredTypeName was not specified, log messages were still parsed into structured object. With this update, parsing of logs is as expected. (LOG-4220)

Logging 5.7.1

This release includes: OpenShift Logging Bug Fix Release 5.7.1.

Bug fixes

  • Before this update, the presence of numerous noisy messages within the Cluster Logging Operator pod logs caused reduced log readability, and increased difficulty in identifying important system events. With this update, the issue is resolved by significantly reducing the noisy messages within Cluster Logging Operator pod logs. (LOG-3482)

  • Before this update, the API server would reset the value for the CollectorSpec.Type field to vector, even when the custom resource used a different value. This update removes the default for the CollectorSpec.Type field to restore the previous behavior. (LOG-4086)

  • Before this update, a time range could not be selected in the OpenShift Container Platform web console by clicking and dragging over the logs histogram. With this update, clicking and dragging can be used to successfully select a time range. (LOG-4501)

  • Before this update, clicking on the Show Resources link in the OpenShift Container Platform web console did not produce any effect. With this update, the issue is resolved by fixing the functionality of the "Show Resources" link to toggle the display of resources for each log entry. (LOG-3218)

Logging 5.7.0

Enhancements

With this update, you can enable logging to detect multi-line exceptions and reassemble them into a single log entry.

To enable logging to detect multi-line exceptions and reassemble them into a single log entry, ensure that the ClusterLogForwarder Custom Resource (CR) contains a detectMultilineErrors field, with a value of true.

Known Issues

None.

Bug fixes

  • Before this update, the nodeSelector attribute for the Gateway component of the LokiStack did not impact node scheduling. With this update, the nodeSelector attribute works as expected. (LOG-3713)