Imported from terminalskills/skills (skills/fluentd/SKILL.md). Install upstream with npx skills add terminalskills/skills --skill fluentd. Copyright stays with the author (Apache-2.0).

Fluentd

Overview

Set up Fluentd as a unified logging layer to collect, filter, transform, and forward logs from applications, containers, and infrastructure to various backends. Covers configuration, plugins, Kubernetes DaemonSet deployment, and buffering strategies.

Instructions

Task A: Basic Fluentd Configuration

<!-- fluent.conf — Collect application logs and forward to Elasticsearch -->
<source>
  @type tail
  path /var/log/app/*.log
  pos_file /var/log/fluentd/app.log.pos
  tag app.logs
  <parse>
    @type json
    time_key timestamp
    time_format %Y-%m-%dT%H:%M:%S.%NZ
  </parse>
</source>

<source>
  @type tail
  path /var/log/nginx/access.log
  pos_file /var/log/fluentd/nginx.log.pos
  tag nginx.access
  <parse>
    @type regexp
    expression /^(?<remote>[^ ]*) - (?<user>[^ ]*) \[(?<time>[^\]]*)\] "(?<method>\S+) (?<path>\S+) (?<protocol>\S+)" (?<status>\d+) (?<size>\d+)/
    time_format %d/%b/%Y:%H:%M:%S %z
  </parse>
</source>

<filter app.logs>
  @type record_transformer
  <record>
    hostname "#{Socket.gethostname}"
    environment production
  </record>
</filter>

<filter app.logs>
  @type grep
  <exclude>
    key level
    pattern /^debug$/
  </exclude>
</filter>

<match app.logs nginx.access>
  @type elasticsearch
  host elasticsearch
  port 9200
  index_name fluentd-${tag}-%Y.%m.%d
  <buffer tag, time>
    @type file
    path /var/log/fluentd/buffer/es
    timekey 1h
    timekey_wait 10m
    flush_interval 30s
    chunk_limit_size 8MB
    total_limit_size 2GB
    retry_max_interval 30
    overflow_action block
  </buffer>
</match>

Task B: Multi-Output Routing

<!-- fluent.conf — Route logs to different destinations based on tag -->
<match error.**>
  @type copy
  <store>
    @type elasticsearch
    host elasticsearch
    port 9200
    index_name errors-%Y.%m.%d
    <buffer time>
      timekey 1d
      flush_interval 10s
    </buffer>
  </store>
  <store>
    @type slack
    webhook_url https://hooks.slack.com/services/T00/B00/XXXX
    channel ops-errors
    message "Error from %s: %s"
    message_keys tag,message
  </store>
</match>

<match audit.**>
  @type s3
  aws_key_id "#{ENV['AWS_ACCESS_KEY_ID']}"
  aws_sec_key "#{ENV['AWS_SECRET_ACCESS_KEY']}"
  s3_bucket audit-logs-production
  s3_region us-east-1
  path logs/%Y/%m/%d/
  <buffer time>
    @type file
    path /var/log/fluentd/buffer/s3
    timekey 3600
    timekey_wait 10m
    chunk_limit_size 256MB
  </buffer>
  <format>
    @type json
  </format>
</match>

<match **>
  @type elasticsearch
  host elasticsearch
  port 9200
  index_name logs-%Y.%m.%d
</match>

Task C: Kubernetes DaemonSet Deployment

# fluentd-daemonset.yml — Fluentd DaemonSet for Kubernetes log collection
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: fluentd
  namespace: logging
  labels:
    app: fluentd
spec:
  selector:
    matchLabels:
      app: fluentd
  template:
    metadata:
      labels:
        app: fluentd
    spec:
      serviceAccountName: fluentd
      tolerations:
        - key: node-role.kubernetes.io/control-plane
          effect: NoSchedule
      containers:
        - name: fluentd
          image: fluent/fluentd-kubernetes-daemonset:v1.16-debian-elasticsearch8-1
          env:
            - name: FLUENT_ELASTICSEARCH_HOST
              value: "elasticsearch.logging.svc.cluster.local"
            - name: FLUENT_ELASTICSEARCH_PORT
              value: "9200"
            - name: FLUENT_ELASTICSEARCH_INDEX_NAME
              value: "k8s-logs"
          resources:
            limits:
              memory: 512Mi
            requests:
              cpu: 100m
              memory: 256Mi
          volumeMounts:
            - name: varlog
              mountPath: /var/log
            - name: containers
              mountPath: /var/lib/docker/containers
              readOnly: true
            - name: config
              mountPath: /fluentd/etc/fluent.conf
              subPath: fluent.conf
      volumes:
        - name: varlog
          hostPath:
            path: /var/log
        - name: containers
          hostPath:
            path: /var/lib/docker/containers
        - name: config
          configMap:
            name: fluentd-config

# fluentd-configmap.yml — Kubernetes-aware Fluentd configuration
apiVersion: v1
kind: ConfigMap
metadata:
  name: fluentd-config
  namespace: logging
data:
  fluent.conf: |
    <source>
      @type tail
      path /var/log/containers/*.log
      pos_file /var/log/fluentd-containers.log.pos
      tag kubernetes.*
      <parse>
        @type cri
      </parse>
    </source>

    <filter kubernetes.**>
      @type kubernetes_metadata
      @id filter_kube_metadata
    </filter>

    <filter kubernetes.**>
      @type grep
      <exclude>
        key $.kubernetes.namespace_name
        pattern /^(kube-system|logging)$/
      </exclude>
    </filter>

    <match kubernetes.**>
      @type elasticsearch
      host "#{ENV['FLUENT_ELASTICSEARCH_HOST']}"
      port "#{ENV['FLUENT_ELASTICSEARCH_PORT']}"
      index_name k8s-${record['$.kubernetes']['namespace_name']}-%Y.%m.%d
      <buffer tag, time>
        @type file
        path /var/log/fluentd/buffer
        timekey 1h
        flush_interval 30s
      </buffer>
    </match>

Task D: Log Parsing and Transformation

<!-- fluent.conf — Parse multiline Java stack traces -->
<source>
  @type tail
  path /var/log/app/java-app.log
  tag java.app
  <parse>
    @type multiline
    format_firstline /^\d{4}-\d{2}-\d{2}/
    format1 /^(?<time>\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}.\d{3}) (?<level>\w+) \[(?<thread>[^\]]+)\] (?<class>[^ ]+) - (?<message>.*)/
  </parse>
</source>

<filter java.app>
  @type record_transformer
  enable_ruby true
  <record>
    severity ${record["level"] == "ERROR" ? "error" : record["level"] == "WARN" ? "warning" : "info"}
    service "java-app"
  </record>
</filter>

Best Practices

• Use file-based buffers in production to prevent log loss during restarts
• Set overflow_action block to apply backpressure instead of dropping logs
• Exclude debug-level logs early in the pipeline to reduce downstream volume
• Use @type copy for critical logs that need multiple destinations
• Monitor Fluentd's own metrics via the @type prometheus plugin
• In Kubernetes, use Fluent Bit as a lightweight forwarder and Fluentd as an aggregator