https://tracker.ceph.com/https://tracker.ceph.com/favicon.ico2015-09-28T18:55:44ZCeph teuthology - Feature #12773: automate identifying infrastructure issues in nightly runshttps://tracker.ceph.com/issues/12773?journal_id=592502015-09-28T18:55:44ZZack Cerza
<ul><li><strong>Assignee</strong> deleted (<del><i>Zack Cerza</i></del>)</li></ul><p>Typed this up a while back and forgot to hit submit:</p>
<p>I spent some time in the last couple days investigating some of the ways we could approach this:</p>
<ul>
<li>An approach similar to John Spray's <code>scrape.py</code></li>
<li>Normalization and grouping of <code>failure_reason</code> values in paddles</li>
<li><a href="http://sentry.ceph.com/" class="external">Sentry</a></li>
</ul>
<p>Ultimately for now Sentry makes the most sense. We had only been using it for Sepia, but I also configured it for Octo.</p>