According to the most recent figures from W3Techs, NGINX powers just over 30% of the websites on the Internet. That’s a lot of websites, and thanks to the growing popularity of NGINX, that number is sure to continue increasing. Because of this popularity, it is highly likely that you are using (or are considering using) NGINX to run your own web applications. Regardless of whether it’s a new or existing installation, monitoring the health of your infrastructure is more important than ever.

This is where Metricly comes in.

Metricly is a full-stack performance monitoring and analytics platform that integrates with just about any process or platform you are likely to use, NGINX included. The beauty of Metricly is that setting up monitors for each integration is incredibly straightforward.

At its core, monitoring server processes with Metricly are an extension of the core Linux agent. While the details of getting the server agent set up are outside the scope of this article, the command itself is so straightforward that it’s worth mentioning. (For more details on how to set up the Linux integration, Metricly has some good documentation to get you started):

sudo N_APIKEY=YOUR_API_KEY bash -c "$(curl -Ls"

Once you’ve set up the Linux agent, getting NGINX monitoring up and running is a snap. The first step to accomplishing this is to configure an internal NGINX status page. To do this, open up the default NGINX site config file (this is usually one of the files in /etc/nginx/sites-enabled/), and add the following block to the server section:

location /nginx_status {\
# turns on nginx stats #\
stub_status on;\
# turns off logging #\
access_log off;\
# sends rest of world to /dev/null #\
deny all;\

Basically, what this block does is create a page at http://localhost/nginx_status that is only accessible locally. If you’re curious about the information this page contains, it is easily accessible by curling the URL from the command line on your server.

root@netuitive-demo:~# curl http://localhost/nginx_status\
Active connections: 1\
server accepts handled requests\
110228 110228 23126\
Reading: 0 Writing: 1 Waiting: 0

When configured, the Metricly agent pushes the information on this page up to your Metricly account, but before that happens, we need to tell the agent about NGINX first. Enabling NGINX monitoring in the Linux agent is as simple as updating one line in one file. To do this, open up the NGINX collector file (found at: /opt/netuitive-agent/conf/collectors/NginxCollector.conf) and change the enabled setting to True.

Because we set up NGINX according to the Metricly documentation, the other settings can be left alone; however, if you deviate from the default configuration, you need to be sure to tell the Linux agent where to find your NGINX status page.

It is important to note here that once everything is configured, it is crucial to restart both your NGINX and Metricly daemons in order to actually send information up to your Metricly account.

After a few minutes, you will start to see NGINX metrics trickle into your Metricly dashboard. The current information shown in Metricly will mirror the information on the status page we configured earlier, but the added benefit to Metricly is that it will also retain the historical process statistics. This allows us to compare information server stats on an hour-by-hour to week-by-week basis.

Monitor NGINX with Metricly: Stats Dashboard NGINX Stats

As you can see, my little demo site has some pretty low statistics, but as a demonstration I spun up a load test using Their free test sends 25 virtual users to a given URL in five minutes, which we can see clearly and immediately on our Metricly dashboard.

Monitor NGINX with Metricly: Connection Spike

NGINX Connection Spike

Sending your NGINX data up to Metricly is a big part of the monitoring puzzle. What you do next is dependent on the needs of your organization. However, Metricly itself offers some great reporting tools that can be configured to alert you and your team of any critical issues.

Ultimately, the most important thing to remember is that monitoring your DevOps infrastructure isn’t optional. Understanding the health of your architecture is critical to supporting the growing needs of your application.

Learn more
About Metricly

Metricly coaches users throughout their cloud journey to organize, plan, analyze, and optimize their public cloud resources.

Try Metricly Free

About the Author