Chapter 2. Logging 6.0

oc explain offers in-depth explanations of all fields associated with a specific object. This includes standard resources like pods and services, as well as more complex entities like statefulsets and custom resources defined by Operators.

To view the documentation for the outputs field of the ClusterLogForwarder custom resource, you can use:

$ oc explain clusterlogforwarders.observability.openshift.io.spec.outputs

This will display detailed information about these fields, including their types, default values, and any associated sub-fields.

The command displays the structure of resource fields in a hierarchical format, clarifying the relationships between different configuration options.

For instance, here’s how you can drill down into the storage configuration for a LokiStack custom resource:

$ oc explain lokistacks.loki.grafana.com
$ oc explain lokistacks.loki.grafana.com.spec
$ oc explain lokistacks.loki.grafana.com.spec.storage
$ oc explain lokistacks.loki.grafana.com.spec.storage.schemas

Each command reveals a deeper level of the resource specification, making the structure clear.

oc explain also indicates the type of each field (such as string, integer, or boolean), allowing you to verify that resource definitions use the correct data types.

For example:

$ oc explain lokistacks.loki.grafana.com.spec.size

This will show that size should be defined using an integer value.

When applicable, the command shows the default values for fields, providing insights into what values will be used if none are explicitly specified.

Again using lokistacks.loki.grafana.com as an example:

$ oc explain lokistacks.spec.template.distributor.replicas

GROUP:      loki.grafana.com
KIND:       LokiStack
VERSION:    v1

FIELD: replicas <integer>

DESCRIPTION:
    Replicas defines the number of replica pods of the component.

Namespace and container inclusions and exclusions have been consolidated into a single field.

apiVersion: "logging.openshift.io/v1"
kind: ClusterLogForwarder
spec:
  inputs:
   - name: application-logs
     type: application
     application:
       namespaces:
       - foo
       - bar
       includes:
       - namespace: my-important
         container: main
       excludes:
       - container: too-verbose

apiVersion: "observability.openshift.io/v1"
kind: ClusterLogForwarder
spec:
  inputs:
   - name: application-logs
     type: application
     application:
       includes:
       - namespace: foo
       - namespace: bar
       - namespace: my-important
         container: main
       excludes:
       - container: too-verbose

Changes to input receivers include:

apiVersion: "logging.openshift.io/v1"
kind: ClusterLogForwarder
spec:
  inputs:
  - name: an-http
    receiver:
      http:
        port: 8443
        format: kubeAPIAudit
  - name: a-syslog
    receiver:
      type: syslog
      syslog:
        port: 9442

apiVersion: "observability.openshift.io/v1"
kind: ClusterLogForwarder
spec:
  inputs:
  - name: an-http
    type: receiver
    receiver:
      type: http
      port: 8443
      http:
        format: kubeAPIAudit
  - name: a-syslog
    type: receiver
    receiver:
      type: syslog
      port: 9442

To use the existing legacy service account logcollector, create the following ClusterRoleBinding:

$ oc adm policy add-cluster-role-to-user collect-application-logs system:serviceaccount:openshift-logging:logcollector
$ oc adm policy add-cluster-role-to-user collect-infrastructure-logs system:serviceaccount:openshift-logging:logcollector

Additionally, create the following ClusterRoleBinding if collecting audit logs:

$ oc adm policy add-cluster-role-to-user collect-audit-logs system:serviceaccount:openshift-logging:logcollector

Inputs are configured in an array under spec.inputs. There are three built-in input types:

Users can define custom inputs of type application that select logs from specific namespaces or using pod labels.

Outputs are configured in an array under spec.outputs. Each output must have a unique name and a type. Supported types are:

Each output type has its own configuration fields.

Pipelines are configured in an array under spec.pipelines. Each pipeline must have a unique name and consists of:

The order of filterRefs matters, as they are applied sequentially. Earlier filters can drop messages that will not be processed by later filters.

Filters are configured in an array under spec.filters. They can match incoming log messages based on the value of structured fields and modify or drop them.

Administrators can configure the following types of filters:

Enables multi-line error detection of container logs.

Log parsers often incorrectly identify separate lines of the same exception as separate exceptions. This leads to extra log entries and an incomplete or inaccurate view of the traced information.

java.lang.NullPointerException: Cannot invoke "String.toString()" because "<param1>" is null
    at testjava.Main.handle(Main.java:47)
    at testjava.Main.printMe(Main.java:19)
    at testjava.Main.main(Main.java:10)

apiVersion: "observability.openshift.io/v1"
kind: ClusterLogForwarder
metadata:
  name: <log_forwarder_name>
  namespace: <log_forwarder_namespace>
spec:
  serviceAccount:
    name: <service_account_name>
  filters:
  - name: <name>
    type: detectMultilineException
  pipelines:
    - inputRefs:
        - <input-name>
      name: <pipeline-name>
      filterRefs:
        - <filter-name>
      outputRefs:
        - <output-name>

When the drop filter is configured, the log collector evaluates log streams according to the filters before forwarding. The collector drops unwanted log records that match the specified configuration.

apiVersion: observability.openshift.io/v1
kind: ClusterLogForwarder
metadata:
# ...
spec:
  serviceAccount:
    name: <service_account_name>
  filters:
  - name: important
    type: drop
    drop:
    - test:
      - field: .message
        notMatches: "(?i)critical|error"
      - field: .level
        matches: "info|warning"
# ...

In addition to including multiple field paths in a single test configuration, you can also include additional tests that are treated as OR checks. In the following example, records are dropped if either test configuration evaluates to true. However, for the second test configuration, both field specs must be true for it to be evaluated to true:

apiVersion: observability.openshift.io/v1
kind: ClusterLogForwarder
metadata:
# ...
spec:
  serviceAccount:
    name: <service_account_name>
  filters:
  - name: important
    type: drop
    drop:
    - test:
      - field: .kubernetes.namespace_name
        matches: "^open"
    - test:
      - field: .log_type
        matches: "application"
      - field: .kubernetes.pod_name
        notMatches: "my-pod"
# ...

OpenShift API servers generate audit events for each API call, detailing the request, response, and the identity of the requester, leading to large volumes of data. The API Audit filter uses rules to enable the exclusion of non-essential events and the reduction of event size, facilitating a more manageable audit trail. Rules are checked in order, and checking stops at the first match. The amount of data that is included in an event is determined by the value of the level field:

The ClusterLogForwarder custom resource (CR) uses the same format as the standard Kubernetes audit policy, while providing the following additional functions:

To disable these defaults, either end your rules list with a rule that has only a level field or add an empty rule.

The ClusterLogForwarder CR audit policy acts in addition to the OpenShift Container Platform audit policy. The ClusterLogForwarder CR audit filter changes what the log collector forwards and provides the ability to filter by verb, user, group, namespace, or resource. You can create multiple filters to send different summaries of the same audit stream to different places. For example, you can send a detailed stream to the local cluster log store and a less detailed stream to a remote site.

apiVersion: observability.openshift.io/v1
kind: ClusterLogForwarder
metadata:
  name: <log_forwarder_name>
  namespace: <log_forwarder_namespace>
spec:
  serviceAccount:
    name: <service_account_name>
  pipelines:
    - name: my-pipeline
      inputRefs: audit 1
      filterRefs: my-policy 2
  filters:
    - name: my-policy
      type: kubeAPIAudit
      kubeAPIAudit:
        # Don't generate audit events for all requests in RequestReceived stage.
        omitStages:
          - "RequestReceived"

        rules:
          # Log pod changes at RequestResponse level
          - level: RequestResponse
            resources:
            - group: ""
              resources: ["pods"]

          # Log "pods/log", "pods/status" at Metadata level
          - level: Metadata
            resources:
            - group: ""
              resources: ["pods/log", "pods/status"]

          # Don't log requests to a configmap called "controller-leader"
          - level: None
            resources:
            - group: ""
              resources: ["configmaps"]
              resourceNames: ["controller-leader"]

          # Don't log watch requests by the "system:kube-proxy" on endpoints or services
          - level: None
            users: ["system:kube-proxy"]
            verbs: ["watch"]
            resources:
            - group: "" # core API group
              resources: ["endpoints", "services"]

          # Don't log authenticated requests to certain non-resource URL paths.
          - level: None
            userGroups: ["system:authenticated"]
            nonResourceURLs:
            - "/api*" # Wildcard matching.
            - "/version"

          # Log the request body of configmap changes in kube-system.
          - level: Request
            resources:
            - group: "" # core API group
              resources: ["configmaps"]
            # This rule only applies to resources in the "kube-system" namespace.
            # The empty string "" can be used to select non-namespaced resources.
            namespaces: ["kube-system"]

          # Log configmap and secret changes in all other namespaces at the Metadata level.
          - level: Metadata
            resources:
            - group: "" # core API group
              resources: ["secrets", "configmaps"]

          # Log all other resources in core and extensions at the Request level.
          - level: Request
            resources:
            - group: "" # core API group
            - group: "extensions" # Version of group should NOT be included.

          # A catch-all rule to log all other requests at the Metadata level.
          - level: Metadata

You can include the application logs based on the label expressions or a matching label key and its values by using the input selector.

When the prune filter is configured, the log collector evaluates log streams according to the filters before forwarding. The collector prunes log records by removing low value fields such as pod annotations.

To apply cluster roles for a user, you must bind an existing cluster role to a specific username.

Cluster roles can be cluster or namespace scoped, depending on which type of role binding you use. When a RoleBinding object is used, as when using the oc adm policy add-role-to-user command, the cluster role only applies to the specified namespace. When a ClusterRoleBinding object is used, as when using the oc adm policy add-cluster-role-to-user command, the cluster role applies to all namespaces in the cluster.

The following example command gives the specified user create, read, update and delete (CRUD) permissions for alerting rules in a specific namespace in the cluster:

$ oc adm policy add-role-to-user alertingrules.loki.grafana.com-v1-admin -n <namespace> <username>

The following command gives the specified user administrator permissions for alerting rules in all namespaces:

$ oc adm policy add-cluster-role-to-user alertingrules.loki.grafana.com-v1-admin <username>

Configurations to ensure Loki’s reliability and efficiency in production.

Workload identity federation enables authentication to cloud-based log stores using short-lived tokens.

The Loki Operator supports setting pod anti-affinity rules to request that pods of the same component are scheduled on different available nodes in the cluster.

Affinity is a property of pods that controls the nodes on which they prefer to be scheduled. Anti-affinity is a property of pods that prevents a pod from being scheduled on a node.

In OpenShift Container Platform, pod affinity and pod anti-affinity allow you to constrain which nodes your pod is eligible to be scheduled on based on the key-value labels on other pods.

The Operator sets default, preferred podAntiAffinity rules for all Loki components, which includes the compactor, distributor, gateway, indexGateway, ingester, querier, queryFrontend, and ruler components.

You can override the preferred podAntiAffinity settings for Loki components by configuring required settings in the requiredDuringSchedulingIgnoredDuringExecution field:

apiVersion: loki.grafana.com/v1
kind: LokiStack
metadata:
  name: logging-loki
  namespace: openshift-logging
spec:
# ...
  template:
    ingester:
      podAntiAffinity:
      # ...
        requiredDuringSchedulingIgnoredDuringExecution: 1
        - labelSelector:
            matchLabels: 2
              app.kubernetes.io/component: ingester
          topologyKey: kubernetes.io/hostname
# ...

When an OpenShift Container Platform cluster is restarted, LokiStack ingestion and the query path continue to operate within the available CPU and memory resources available for the node. This means that there is no downtime for the LokiStack during OpenShift Container Platform cluster updates. This behavior is achieved by using PodDisruptionBudget resources. The Loki Operator provisions PodDisruptionBudget resources for Loki, which determine the minimum number of pods that must be available per component to ensure normal operations under certain conditions.

Specialized configurations for high availability, scalability, and error handling.

The Loki Operator offers support for zone-aware data replication through pod topology spread constraints. Enabling this feature enhances reliability and safeguards against log loss in the event of a single zone failure. When configuring the deployment size as 1x.extra-small, 1x.small, or 1x.medium, the replication.factor field is automatically set to 2.

To ensure proper replication, you need to have at least as many availability zones as the replication factor specifies. While it is possible to have more availability zones than the replication factor, having fewer zones can lead to write failures. Each zone should host an equal number of instances for optimal operation.

apiVersion: loki.grafana.com/v1
kind: LokiStack
metadata:
 name: logging-loki
 namespace: openshift-logging
spec:
 replicationFactor: 2 1
 replication:
   factor: 2 2
   zones:
   -  maxSkew: 1 3
      topologyKey: topology.kubernetes.io/zone 4

In OpenShift Container Platform a zone failure happens when specific availability zone resources become inaccessible. Availability zones are isolated areas within a cloud provider’s data center, aimed at enhancing redundancy and fault tolerance. If your OpenShift Container Platform cluster is not configured to handle this, a zone failure can lead to service or data loss.

Loki pods are part of a StatefulSet, and they come with Persistent Volume Claims (PVCs) provisioned by a StorageClass object. Each Loki pod and its PVCs reside in the same zone. When a zone failure occurs in a cluster, the StatefulSet controller automatically attempts to recover the affected pods in the failed zone.

The StatefulSet controller automatically attempts to reschedule pods in a failed zone. Because the associated PVCs are also in the failed zone, automatic rescheduling to a different zone does not work. You must manually delete the PVCs in the failed zone to allow successful re-creation of the stateful Loki Pod and its provisioned PVC in the new zone.

If the Log Forwarder API forwards a large block of messages that exceeds the rate limit to Loki, Loki generates rate limit (429) errors.

These errors can occur during normal operation. For example, when adding the logging to a cluster that already has some logs, rate limit errors might occur while the logging tries to ingest all of the existing log entries. In this case, if the rate of addition of new logs is less than the total rate limit, the historical data is eventually ingested, and the rate limit errors are resolved without requiring user intervention.

In cases where the rate limit errors continue to occur, you can fix the issue by modifying the LokiStack custom resource (CR).