Tech Blog: Connecting Elasticsearch logs to Grafana and Kibana
By Tomislav Levanić on July 18, 2022
What is the point in collecting logs and metrics if you don’t use them? In this blog post, we will build upon our previous blog post and connect Fluent Bit log collectors to Elasticsearch along with a basic setup and comparison of Kibana and Grafana, tools often used for visualizing logs
This blog post is a part of the Velebit AI Tech Blog series where we discuss good practices for a scalable and robust production deployment.
This post is heavily dependent on the previous blog post in this series so you can always go back and check it out if you are feeling a little lost with the Fluent Bit aspect of the configuration. We are going to modify all configuration files except log generator service from ./service.yml. Just like the last time, you can find configuration files on GitHub. Now you should be ready for the rest of the blog.
Store logs to Elasticsearch
To visualize logs first we need to store historic values somewhere. We opted for Elasticsearch as we have experience with it and it has served us well for years. This means that along with Fluent Bit service that collects logs, we now also need an Elasticsearch instance. Let’s start by creating logging.yml with both of these configurations
somewhere in the services section, along with updating the volumes section to
in order to preserve Grafana configuration and user data. Note that Kibana uses the Elasticsearch instance for saving its configuration which is why this step is not needed for Kibana.
To connect Grafana to Elasticsearch, you simply need to add a new data source and fill in:
Data Source settings
Time field name
Table 2. Grafana Elasticsearch Data Source settings
After this, press the Save & Test button and you are ready to go!
Which one to choose?
Kibana is a free and open user interface that lets you visualize your Elasticsearch data and navigate the Elastic Stack. It integrates seamlessly with Elasticsearch and provides an excellent experience for exploring both logs and various metrics extracted from logs. Elastic also offers the X-Pack license to provide even more capabilities. One example would be alerting, which does not come together with Elastic without X-Pack.
Grafana is in general more oriented towards metrics than the logs. Even though you can inspect logs and stack traces, the experience you get from Kibana is simply superior. The most important advantage Grafana has over Kibana is that even the free version of Grafana allows you to query, visualize, alert on and understand your metrics no matter where they are stored. This opens up a lot of possibilities. For example, we used Grafana together with InfluxDB to build a historical overview of official Covid19-related data for Croatia. You can see the board in action over at https://covid19hr.velebit.ai
Still can’t choose? Use both of them. That is exactly what we at Velebit AI do as well! We prefer Grafana for metrics and alerting, while we use Kibana once something goes wrong and we need to delve into the application logs.
Configure Fluent Bit to work with Elasticsearch
Finally, we must configure Fluent Bit to send the logs it collects to the Elasticsearch. We can do this by simply adding another OUTPUT section at the end of the config.conf file from the previous blog post.
# send logs to elasticsearch
This would make the final file look like this
# import parsers defined in a parsers file
# collect docker logs using fluend logging driver
# make SERVICE_NAME key lowercase
Rename SERVICE_NAME service_name
# try parsing log as json and lift its keys to the first-level
# send logs to stdout
# send logs to elasticsearch
After modifying all the necessary files, we can start the defined services using
docker-compose -f logging.yml up -d
docker-compose -f service.yml up -d
An easy way to check whether the setup is working or not is to make an HTTP request directly to Elasticsearch, for example
which should return a list of entries in the specified index. This means you are ready to go and explore your logs in Grafana and Kibana!
In this blog post we used logs created by dummy log-generating service defined in ./service.yml file. In production environments you will have to collect logs from various sources and you can do that easily as long as Fluent Bit can parse their logs. We decided to use JSON log format for everything as it simplifies a lot of things. If you want to learn more about this you can check out our previous blog posts where we discuss how we solved the problems with formating uwsgi and Python logs as JSON.
Finally, complete cleanup can be executed using
docker-compose -f service.yml down --volumes
docker-compose -f logging.yml down
We may want to develop a model that solves the detection task, and simultaneously predicts some additional attributes that can be related to the whole image or to the detected instances. This blog post demonstrates how to implement that in the MMDetection framework.
Filtering by color is a feature we often use on fashion marketplaces, e-commerce sites, and any other sites where we buy clothes. Color is also important when buying furniture, and antiques, or searching for stock images.
To deliver a smooth user experience, it's essential to stay current with technology in the realm of UX design. In this blog post, we'll look at how AI is becoming increasingly indispensable in web app development to give users a more tailored experience.
We build AI for your needs.
Meet our highly experienced team, who loves to build AI and design its surroundings to incorporate it in your business. Find out how much you can benefit from our fair and open approach.