CINXE.COM
logsnarf • Consumes Logs → Emits Metrics
<!DOCTYPE html><html><head><meta content="heroku logs metrics time-series devops" name="keywords" /><meta content="logsnarf is a tool for Heroku users to extract their Heroku metrics data and load it in a time-series database" name="description" /><meta content="utf-8" name="charset" /><meta content="width=device-width, initial-scale=1" name="viewport" /><link href="/images/logo.svg" rel="icon" type="image/svg" /><link href="/stylesheets/site.css" rel="stylesheet" /><script src="/javascripts/site.js"></script><title>logsnarf • Consumes Logs → Emits Metrics</title></head></html><body><div><nav><a class="brand" href="/"><img src="/images/logo.svg" height="18px" alt="Logo" />LogSnarf </a><a class="faq" href="/faq.html">FAQ </a></nav></div><main><section class="infobox"><div class="content"><img src="/images/logo.svg" height="64" class="logo" alt="Logo" /><div class="intro"><p><strong><span class="brand">LogSnarf</span> is the simplest way to get metrics from your Heroku apps.</strong> Heroku captures many useful metrics about your dynos and hosted databases, but they're only exposed in logs or the Heroku UI. Add <span class="brand">LogSnarf</span> as a <em>Heroku Log Drain</em>, and it will extract all the metrics it find and push them to your time-series database.</p> </div><p><span class="brand">LogSnarf</span> is currently in private beta (<a href="/faq.html#beta">what does beta mean?</a>) while we test things out and estimate the load. If you're interesting in helping us and want try it out, <a href="mailto:beta@logsnarf.com">send us an email</a>!</p> </div></section><section class="content"><div class="grid"><div class="info"><h2>Parse Metrics From Logs</h2><p>Heroku collects several metrics about your dynos and hosted databases, but does not expose them except via the log files and their own dashboard tool. <span class="brand">LogSnarf</span> can extract these metrics from the Heroku logs:</p> <ul> <li>Worker/Web Dynos — Memory & Load Avg</li> <li>PostgreSQL — Memory, Load, IOPS, Connections, Cache Hit Rate</li> <li>Redis — Memory, Load, IOPS, Connections, Cache Hit Rate</li> </ul> </div><div class="preview"><a href="/images/dashboard-dynos.png" target="_blank" rel="noopener"><img src="/images/dashboard-dynos.png" alt="Dashboard dynos" /></a><a href="/images/dashboard-dyno.png" target="_blank" rel="noopener"><img src="/images/dashboard-dyno.png" alt="Dashboard dyno" /></a></div></div><div class="grid"><div class="info"><h2>Push Metrics to a Time-Series Database</h2><p><span class="brand">LogSnarf</span> integrates with your existing metrics store.</p> <ul> <li><a href="https://www.influxdata.com/">InfluxDB</a> (1.x and 2.x)</li> <li><a href="https://prometheus.io/">Prometheus</a> <sup>†</sup></li> <li><a href="https://www.timescale.com/">TimescaleDB</a> <sup>†</sup></li> <li><a href="http://opentsdb.net">OpenTSDB</a> <sup>†</sup></li> </ul> <small> <sup>†</sup>: <span class="brand">LogSnarf</span> currently only supports InfluxDB. Support for other databases will be added based on customer need. </small></div><div class="preview"><a href="/images/dashboard-postgresql.png" target="_blank" rel="noopener"><img src="/images/dashboard-postgresql.png" alt="Dashboard postgresql" /></a></div></div><div class="grid"><div class="info"><h2>Add Metrics to your Dashboard</h2><p>The metrics that <span class="brand">LogSnarf</span> adds to your time-series DB can be graphed or alerted on like any other metrics.</p> </div><div class="preview"><a href="/images/dashboard-postgresql.png" target="_blank" rel="noopener"><img src="/images/dashboard-postgresql.png" alt="Dashboard postgresql" /></a></div></div></section></main></body>