The ELK Stack, traditionally comprised of Elasticsearch, Logstash, and Kibana, now also includes a fourth element – Beats, a family of log shippers for different use cases and sets of data. To reflect this structural and functional change, the stack is being renamed as the Elastic Stack.
Beats include a variety of different log shippers. In a previous tutorial, we described how to work with Filebeat for shipping log files into the stack. Despite some structural similarities, Metricbeat is a bit different, and this tutorial will outline the differences as well as how to work with this shipper.
What is Metricbeat?
Metricbeat evolved out of Topbeat and like the other beats, is built upon Libbeat – a Go framework. As its name implies, Metricbeat is a lightweight shipper that collects and ships various system and service metrics to a specified output destination.
Metricbeat is installed on the different servers in your environment and used for monitoring their performance, as well as that of different external services running on them. For example, you can use Metricbeat to monitor and analyze system CPU, memory and load. In Dockerized environments, Metricbeat can be installed on a host for monitoring container performance metrics.
Metricbeat is usually configured to ship the data directly to an Elasticsearch deployment. To gain architectural flexibility though, Metricbeat can also be configured to ship to a messaging or queuing buffer such as Kafka or Redis.
Metricbeat can be downloaded and installed using a variety of different methods and on a variety of supported platforms. Of course, you will need to have a running ELK Stack to be able to ship and store the collected metrics. I will outline two methods, using Apt and Docker, but you can refer to the official docs for more options.
Install Metricbeat using Apt
For an easier way of updating to a newer version, and depending on your Linux distro, you can use Apt or Yum to install Metricbeat from Elastic’s repositories:
First, you need to add Elastic’s signing key so that the downloaded package can be verified (skip this step if you’ve already installed packages from Elastic):
wget -qO - https://artifacts.elastic.co/GPG-KEY-elasticsearch | sudo apt-key
The next step is to add the repository definition to your system:
echo "deb https://artifacts.elastic.co/packages/6.x/apt stable main" | sudo
tee -a /etc/apt/sources.list.d/elastic-6.x.list
All that’s left to do is to update your repositories and install Metricbeat:
sudo apt-get update && sudo apt-get install metricbeat
Install Metricbeat on Docker
If you’re running Docker, you can install Metricbeat as a container on your host and configure it to collect either host metrics or container metrics.
To install Metricbeat as a container, pull Elastic’s Metricbeat image:
docker pull docker.elastic.co/beats/metricbeat:6.2.3
Before we run Metricbeat, we need to understand the basics of how to configure it, first and foremost what configuration options are available within the main configuration file.
Metricbeat configuration file
Metricbeat is configured using a YAML configuration file. On Linux, this file is located at: /etc/metricbeat/metricbeat.yml. On Docker, you will find it at: /usr/share/metricbeat/metricbeat.yml.
YAML is syntax sensitive. You cannot, for example, use tabs for spacing. There are a number of additional best practices that will help you avoid mistakes in this Musing in YAML article.
Metricbeat modules and metricsets
Metricbeat modules contain service-specific collection and connection definitions. They define what specific metrics to collect from the service, the frequency in which to collect them, and how to connect to it. Modules, in turn, are comprised of one or multiple metricsets — literally, a set of related metrics to be collected and shipped.
Metricbeat supports a growing number of modules, such as System, Apache, nginx, MySQL and Docker, to name just the popular and more commonly used modules. By default, Metricbeat is configured to use the system module which collects server system metrics, such as CPU and memory usage, network, and so forth.
Here is an example of a Metricbeat configuration file that has two modules defined, the System and Apache modules:
- module: system
- module: apache
Metricbeat configuration settings
For each module, define a list of module-specific metricsets. We cannot define a metricset that is not supported by the module used.
Using the period setting, we define the frequency at which we want Metricbeat to fetch the metrics.
The hosts setting is optional, and it defines the host, or hosts, we want Metricset to fetch the metrics from. You can also use fields and tags to add custom fields and tags to the metricset events.
Another optional setting is processors, which can be used to apply different changes to the data collected by Metricbeat. For example, you can use processors to drop specific fields, drop specific events, add metadata and more.There are two main methods for enabling modules — you can enable them in the modules.d directory in which they are installed (/etc/metricbeat/modules.d) or as shown in the example above from within the configuration file. (enabled: true).
Configuring Metricbeat to monitor Docker
Using the Docker module, Metricbeat can be configured to ship a bunch of useful information about Docker containers running on the host:
- module: docker
metricsets: ["container", "cpu", "diskio", "healthcheck", "info", "memory", "network"]
This section in the Metricbeat configuration file defines where you want to ship the metrics to. As specified above, in most cases you will make do with shipping the events directly into Elasticsearch, but Metricbeat can be configured to ship the data to other destinations.
For more advanced processing and data enhancement, you can define Logstash as your output (remember to use the beats input plugin in Logstash to collect the data from metricbeat). Kafka and Redis are also supported as output destinations, for more complex and traffic-heavy pipelines.
For forwarding logs to Elasticsearch:
For forwarding logs to Logstash:
Now that we understand the basics of configuring Metricbeat, it’s time to actually run the program.
Before we do so, however, it’s a good practice to test your configuration file. To do this, access the directory where Metricbeat is installed, and run Metricbeat with the test subcommand:
sudo metricbeat -c /etc/metricbeat/metricbeat.yml test config
Depending on how you installed Metricbeat, enter the following commands to start Metricbeat.
Start the Metricbeat service with:
sudo service metricbeat start
Run the Metricbeat container with defining bind-mounting to your configuration file (you can of course do the same thing by building your own image from a Dockerfile and running it). Be sure you have the correct permissions to connect to the Docker daemon:
sudo docker run -v
If all runs as expected, you should be seeing a new index created in Elasticsearch:
curl -XGET 'localhost:9200/_cat/indices?v&pretty'
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
yellow open metricbeat-6.2.2-2018.04.16 YePjpOgqQR-pU_SuKrm5vw 1 1 2989 0 1.1mb 1.1mb
Define the new index pattern in Kibana to start the analysis.
Shipping to Logz.io
By making a few adjustments to your Metricbeat configuration file, metrics can be sent to the Logz.io ELK Stack for analysis and visualization.
First, you will need to download an SSL certificate to use encryption:
sudo mkdir -p /etc/pki/tls/certs
sudo cp COMODORSADomainValidationSecureServerCA.crt /etc/pki/tls/certs/
Next, retrieve your Logz.io account token from the UI (under Settings → Manage Accounts).
Finally, to ship the data to Logz.io, your Metricbeat configuration file needs to look as follows:
- module: system
Logz.io provides ready-made dashboards for monitoring the data shipped by Metricbeat. To install it, simply open ELK Apps, Logz.io’s library of pre-made dashboards and visualizations and search for ‘metricbeat’.
In this case, we installed the Docker Monitoring dashboard:
Metricbeat is an extremely easy-to-use, efficient and reliable metric shipper for monitoring your system and the processes running on it. Like the other beats in the family — Filebeat, Winlogbeat, Packetbeat, Auditbeat and heartbeat, Metricbeat leaves a light resource footprint and can be installed and started relatively quickly.
Like in any software, there are a some quirks, mainly related to issues with configuring YAML syntax. Tips on configuring YAML configuration files can be found in the Musings in YAML article.