How can I collect HTTP response codes from Nginx?

The Datadog integration collects HTTP response codes and more for NGINX Plus, however, to collect similar metrics for non-Plus you'll need to parse the access logs.

The following script was provided by a Datadog customer (Credit: Ganesh Swami), it shows how it's possible to collect 5XX codes:

This article is intended to serve as an example for those that are interested in parsing the NGINX access.log using the parser built into the Datadog agent.

Configuration steps

First, let's copy the code from the gist in a file that's accessible by the agent, here we'll use: /usr/share/datadog/agent/dogstream/

Be certain this file is accessible by the agent:

$ sudo chown dd-agent /usr/share/datadog/agent/dogstream/
$ sudo ls -al /usr/share/datadog/agent/dogstream/
-rw-r--r-- 1 dd-agent root 2313 Oct 22 14:02 /usr/share/datadog/agent/dogstream/

Now we must edit NGINX's log format to be compatible with this parser, more information about this here:

$ grep 'log_format' -A 3 /etc/nginx/nginx.conf
    log_format time_log '$time_local "$request" S=$status $bytes_sent T=$request_time R=$http_x_forwarded_for';
    access_log /var/log/nginx/access.log time_log ;
    error_log /var/log/nginx/error.log;

Then we restart the service, generate a request using curl and check the log file to make sure the log format has changed:

$ sudo service nginx restart
 * Restarting nginx nginx    [ OK ]
$ curl localhost > /dev/null
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   612  100   612    0     0  57357      0 --:--:-- --:--:-- --:--:-- 61200
$ cat /var/log/nginx/access.log - - [22/Oct/2015:13:50:23 +0000] "GET / HTTP/1.1" 200 612 "-" "curl/7.35.0"
22/Oct/2015:13:54:43 +0000 "GET / HTTP/1.1" S=200 858 T=0.000 R=-

The dd-agent user may be unable to read the log file. You need to add the dd-agent user to the adm group:

$ sudo usermod -a -G adm dd-agent

You can check that the dd-agent user has access now.

$ ls -al /var/log/nginx/access.log
 -rw-r--r-- 1 root root 1044 Oct 22 15:24 /var/log/nginx/access.log
$ sudo -u dd-agent wc -l /var/log/nginx/access.log
 2 /var/log/nginx/access.log

We can now configure the agent to use the parser's parse function on the access.log file:

$ grep '^dogstream' /etc/dd-agent/datadog.conf
dogstreams: /var/log/nginx/access.log:/usr/share/datadog/agent/dogstream/

Let's restart the agent and check the /var/log/datadog/collector.log file to make sure that the log file is beeing read:

$ cat /var/log/datadog/collector.log
2015-10-22 15:34:15 UTC | INFO | dd.collector | collector( | Exiting. Bye bye.
2015-10-22 15:34:17 UTC | INFO | dd.collector | utils.pidfile( | Pid file is: /opt/datadog-agent/run/
2015-10-22 15:34:17 UTC | INFO | dd.collector | collector( | Agent version 5.5.1
2015-10-22 15:34:17 UTC | INFO | dd.collector | checks.collector( | Instantiating function-based dogstream
2015-10-22 15:34:17 UTC | INFO | dd.collector | checks.collector( | dogstream: parsing /var/log/nginx/access.log with <function parse at 0x7fbef0250488> (requested /usr/share/datadog/agent/dogstream/
2015-10-22 15:34:17 UTC | INFO | dd.collector | checks.collector( | Dogstream parsers: [<checks.datadog.Dogstream object at 0x7fbef0237d10>]

Everything looks good, so let's generate some HTTP requests using the curl command seen above.

Visualizing metrics

As you can see from the code, the parser will generate a gauge metric for each request. Let's make sure that the metric was received by Datadog, by browsing the Metric Summary page:

We can now plot the metric in Metric Explorer:

Have more questions? Submit a request


Please sign in to leave a comment.
Powered by Zendesk