Docker Compose

This article explains how to collect Docker logs and propagate them to EFK (Elasticsearch + Fluentd + Kibana) stack. The example uses Docker Compose for setting up multiple containers.

Kibana

Elasticsearch had been an open-source search engine known for its ease of use. Kibana had been an open-source Web UI that makes Elasticsearch user-friendly for marketers, engineers and data scientists alike.

NOTE: Since v7.11, These products are distributed under non open-source license (Dual licensed under Server Side Public License and Elastic License)

By combining these three tools EFK (Elasticsearch + Fluentd + Kibana) we get a scalable, flexible, easy to use log collection and analytics pipeline. In this article, we will set up four (4) containers, each includes:

• Apache HTTP Server

• Fluentd

• Elasticsearch

• Kibana

All the logs of httpd will be ingested into Elasticsearch + Kibana, via Fluentd.

Prerequisites: Docker

Please download and install Docker / Docker Compose. Well, that's it :)

• Docker Installation

Step 0: Create docker-compose.yml

Create docker-compose.yml for Docker Compose. Docker Compose is a tool for defining and running multi-container Docker applications.

With the YAML file below, you can create and start all the services (in this case, Apache, Fluentd, Elasticsearch, Kibana) by one command:

services:
  web:
    image: httpd
    ports:
      - "8080:80"
    depends_on:
      - fluentd
    logging:
      driver: "fluentd"
      options:
        fluentd-address: localhost:24224
        tag: httpd.access

  fluentd:
    build: ./fluentd
    volumes:
      - ./fluentd/conf:/fluentd/etc
    depends_on:
      # Launch fluentd after that elasticsearch is ready to connect
      elasticsearch:
        condition: service_healthy
    ports:
      - "24224:24224"
      - "24224:24224/udp"

  elasticsearch:
    image: docker.elastic.co/elasticsearch/elasticsearch:8.17.1
    container_name: elasticsearch
    hostname: elasticsearch
    environment:
      - discovery.type=single-node
      - xpack.security.enabled=false # Disable security for testing
    healthcheck:
      # Check whether service is ready
      test: ["CMD", "curl", "-f", "http://localhost:9200/_cluster/health"]
      interval: 10s
      retries: 5
      timeout: 5s
    ports:
      - 9200:9200

  kibana:
    image: docker.elastic.co/kibana/kibana:8.17.1
    depends_on:
      # Launch fluentd after that elasticsearch is ready to connect
      elasticsearch:
        condition: service_healthy
    ports:
      - "5601:5601"

The logging section (check Docker Compose documentation) of web container specifies Docker Fluentd Logging Driver as a default container logging driver. All the logs from the web container will automatically be forwarded to host:port specified by fluentd-address.

Step 1: Create Fluentd Image with your Config + Plugin

Create fluentd/Dockerfile with the following content using the Fluentd official Docker image; and then, install the Elasticsearch plugin:

# fluentd/Dockerfile

FROM fluent/fluentd:edge-debian
USER root

# To connect to docker.elastic.co/elasticsearch/elasticsearch:8.x, it requires elasticsearch v8 gem
# Ref. https://github.com/elastic/elasticsearch-ruby/blob/main/README.md#compatibility
RUN ["gem", "install", "elasticsearch", "--no-document", "--version", "8.19.0"]

RUN ["gem", "install", "fluent-plugin-elasticsearch", "--no-document", "--version", "5.4.3"]
USER fluent

Then, create the Fluentd configuration file fluentd/conf/fluent.conf. The forward input plugin receives logs from the Docker logging driver and elasticsearch output plugin forwards these logs to Elasticsearch.

# fluentd/conf/fluent.conf

<source>
  @type forward
  port 24224
  bind 0.0.0.0
</source>

<match *.**>
  @type copy

  <store>
    @type elasticsearch
    host elasticsearch
    port 9200
    logstash_format true
    logstash_prefix fluentd
    logstash_dateformat %Y%m%d
    include_tag_key true
    type_name access_log
    tag_key @log_name
    flush_interval 1s
  </store>

  <store>
    @type stdout
  </store>
</match>

NOTE: The detail of used parameters for @type elasticsearch, see Elasticsearch parameters section and fluent-plugin-elasticsearch furthermore.

Step 2: Start the Containers

Let's start the containers:

$ docker compose up --detach

Use docker ps command to verify that the four (4) containers are up and running:

$ docker ps
CONTAINER ID   IMAGE                                                  COMMAND                   CREATED          STATUS                    PORTS                                                                                                    NAMES
7a489886d856   httpd                                                  "httpd-foreground"        36 seconds ago   Up 14 seconds             0.0.0.0:8080->80/tcp, [::]:8080->80/tcp                                                                  fluentd-elastic-kibana-web-1
36ded62da733   fluentd-elastic-kibana-fluentd                         "tini -- /bin/entryp…"    36 seconds ago   Up 15 seconds             5140/tcp, 0.0.0.0:24224->24224/tcp, 0.0.0.0:24224->24224/udp, :::24224->24224/tcp, :::24224->24224/udp   fluentd-elastic-kibana-fluentd-1
254b7692966f   docker.elastic.co/kibana/kibana:8.17.1                 "/bin/tini -- /usr/l…"    36 seconds ago   Up 15 seconds             0.0.0.0:5601->5601/tcp, :::5601->5601/tcp                                                                fluentd-elastic-kibana-kibana-1
187d3e5c2e08   docker.elastic.co/elasticsearch/elasticsearch:8.17.1   "/bin/tini -- /usr/l…"    37 seconds ago   Up 35 seconds (healthy)   0.0.0.0:9200->9200/tcp, :::9200->9200/tcp, 9300/tcp                                                      elasticsearch

Step 3: Generate httpd Access Logs

Use curl command to generate some access logs like this:

$ curl http://localhost:8080/
<html><body><h1>It works!</h1></body></html>

Step 4: Confirm Logs from Kibana

Browse to http://localhost:5601/app/discover#/ and create data view.

Specify fluentd-* to Index pattern and click Save data view to Kibana.

Then, go to Discover tab to check the logs. As you can see, logs are properly collected into the Elasticsearch + Kibana, via Fluentd.

Kibana Discover

Learn More

• Fluentd: Architecture

• Fluentd: Get Started

• Downloading Fluentd

If this article is incorrect or outdated, or omits critical information, please let us know. Fluentd is an open-source project under Cloud Native Computing Foundation (CNCF). All components are available under the Apache 2 License.