healthchecks/templates/docs/attaching_logs.html

<h1>Attaching Logs</h1>
<p>SITE_NAME ping endpoints accept HTTP HEAD, GET and POST request methods.</p>
<p>When using HTTP POST, <strong>you can include an arbitrary payload in the request body</strong>.
If the request body looks like a UTF-8 string, SITE_NAME will log the
first 10 kilobytes (10 000 bytes) of the request body, so that you can inspect
it later.</p>
<h2>Logging Command Output</h2>
<p>In this example, we run <code>certbot renew</code>, capture its output (both the stdout
and stderr streams), and submit the captured output to SITE_NAME:</p>
<div class="highlight"><pre><span></span><code><span class="ch">#!/bin/sh</span>

<span class="nv">m</span><span class="o">=</span><span class="k">$(</span>/usr/bin/certbot renew <span class="m">2</span>&gt;<span class="p">&amp;</span><span class="m">1</span><span class="k">)</span>
curl -fsS -m <span class="m">10</span> --retry <span class="m">5</span> --data-raw <span class="s2">&quot;</span><span class="nv">$m</span><span class="s2">&quot;</span> PING_URL
</code></pre></div>

<h2>In Combination with the <code>/fail</code> and <code>/{exit-status}</code> Endpoints</h2>
<p>We can extend the previous example and signal either success or failure
depending on the exit code:</p>
<div class="highlight"><pre><span></span><code><span class="ch">#!/bin/sh</span>

<span class="nv">m</span><span class="o">=</span><span class="k">$(</span>/usr/bin/certbot renew <span class="m">2</span>&gt;<span class="p">&amp;</span><span class="m">1</span><span class="k">)</span>
curl -fsS -m <span class="m">10</span> --retry <span class="m">5</span> --data-raw <span class="s2">&quot;</span><span class="nv">$m</span><span class="s2">&quot;</span> PING_URL/<span class="nv">$?</span>
</code></pre></div>

<h2>Using Runitor</h2>
<p><a href="https://github.com/bdd/runitor">Runitor</a> is a third party utility that runs the
supplied command, captures its output and reports to SITE_NAME.
It also measures the execution time and retries HTTP requests on transient errors.
Best of all, the syntax is simple and clean:</p>
<div class="highlight"><pre><span></span><code>runitor -uuid your-uuid-here -- /usr/bin/certbot renew
</code></pre></div>

<h2>Handling More Than 10KB of Logs</h2>
<p>While SITE_NAME can store a small amount of logs in a pinch, it is not specifically
designed for that. If you run into the issue of logs getting cut off, consider
the following options:</p>
<ul>
<li>See if the logs can be made less verbose. For example, if you have a batch job
that outputs a line of text per item processed, perhaps it can output a summary with
the totals instead.</li>
<li>If the important content is usually at the end, submit the <strong>last 10KB</strong> instead
of the first. Here is an example that submits the last 10KB of <code>dmesg</code> output:</li>
</ul>
<div class="highlight"><pre><span></span><code><span class="ch">#!/bin/sh</span>

<span class="nv">m</span><span class="o">=</span><span class="k">$(</span>dmesg <span class="p">|</span> tail --bytes<span class="o">=</span><span class="m">10000</span><span class="k">)</span>
curl -fsS -m <span class="m">10</span> --retry <span class="m">5</span> --data-raw <span class="s2">&quot;</span><span class="nv">$m</span><span class="s2">&quot;</span> PING_URL
</code></pre></div>

<ul>
<li>Finally, if it is critical to capture the entire log output,
consider using a dedicated log aggregation service for capturing the logs.</li>
</ul>