In this tutorial we will cover the basic EFK setup which is Fluent Bit, Elasticsearch and Kibana for log aggregation/collection, parsing the logs, storing the logs in Elasticsearch and finally visualizing the logs using the Kibana UI.
These days, with growing focus on distributed and microservice based architecture, where different services may run on different virtual machines or servers, it becomes crucial to have some setup in place for log collection, to have all the logs collected and stored at one place and if you want to do analysis on the collected logs having Kibana UI or Grafana UI is of great benefit.
Fluent Bit is a data collection service, Elasticsearch is a service to store data in JSON format and Kibana is UI service which can be configured to stream data from Elasticsearch service. We will see how we can do the basic installation of all these services on a Linux machine on a non Kubernetes environment.
We have a separate tutorial covering installation steps of Fluent Bit. To configure Fluent Bit, we will have to setup Input and Output configuration so that we can read logs from our Application log file, or in case of multiple applications, we can also configure Fluent Bit to tail logs from multiple log files.
Once you have installed Fluent Bit on your Linux Machine, you will find its configuraiton files in /etc/td-agent-bit/ directory. You will have a td-agent-bit.conf configuration file in which we can provide Input plugin and Output plugin.
In our setup, we will use the Input tail plugin to tail data from log files, and we will have the Elasticsearch output plugin to push the data into Elasticsearch, creating an index on which the data will be stored.
Following should be the Input configuration:
[INPUT] Name tail Tag mylogs Path /path/to/files/*.log Mem_Buf_Limit 5MB Skip_Long_Lines Off Refresh_Interval 10
and the output configuration:
[OUTPUT] Name es Match * Index mylogs Host localhost Port 9200 Logstash_Format Off
In the above configuration we have specified the Hostname of the server on which elastic search is running, if you setup all the three service on the same machine, you can use localhost too, else change it to the hostname of your machine and the port on which the Elasticsearch runs is 9200 by default.
Restart the Fluent bit service using the following command:
sudo service td-agent-bit restart
Installing Elasticsearch is very easy. Download the zip file for Elasticsearch from here: https://www.elastic.co/downloads/elasticsearch
Unzip the zip file downloaded. Then all you have to do is go to config folder and update the configuration file for elasticsearch in elasticsearch.yml file. All you have to do is provide the hostname for the server if your host has a domain name.
To start the service, go to the bin/ directory and run the elasticsearch shell file. Run the following command:
nohup ./elasticsearch &
This will run the service in background.
To install kibana, download Kibana from here: https://www.elastic.co/downloads/kibana
Make sure you download the same version number of Kibana as of Elasticsearch. Unzip the zip file downloaded. Go to the kibana folder, in the config directory and update the kibana.yml file and change the hostname from localhost to your server's hostname if required.
Then go to bin/ directory and run the kibana shell script, using the following command:
nohup ./kibana &
This will run the service in the background. Once up, the Kibana service will be available at http://localhost:5601/app/kibana URL.
If all is configured well, your Fluent Bit service should start reading logs from the log files and push it to Elasticsearch and you should be able to see the logs in Kibana.
To see if Fluent Bit is working fine and is creating index in Elasticsearch you can run the following command to see the indices created in Elasticsearch:
curl -XGET 'localhost:9200/_cat/indices?v&pretty'
To see the logs in Kibana UI, you should create a new Index pattern for your index which is mylogs, and then use the Kibana UI to to see logs for that index.
If you face any confusion with the following setup, feel free to comment down below.