https://blog.devbar.app/posts/Recent content in Posts on DevBar BlogHugoenThu, 30 Apr 2026 00:00:00 +0000https://blog.devbar.app/posts/why-we-built-devbar/Thu, 30 Apr 2026 00:00:00 +0000https://blog.devbar.app/posts/why-we-built-devbar/<p>It was 2 AM during a production incident. A payment service was returning 503s. I had PagerDuty open in one tab, the relevant GitHub PR in another, a Datadog dashboard in a third, Linear for the incident ticket, and Slack somewhere underneath all of it. I spent four minutes — four actual minutes — just trying to find the deployment that caused the regression.</p> <p>The code fix took thirty seconds. Finding it took four minutes.</p>https://blog.devbar.app/posts/building-real-time-alerts-without-polling/Tue, 28 Apr 2026 00:00:00 +0000https://blog.devbar.app/posts/building-real-time-alerts-without-polling/<p>The first version of DevBar polled every integration on a 60-second interval. It worked, but 60 seconds is an eternity during a production incident. By the time the menu bar updated to show a new PagerDuty alert, the on-call engineer had already gotten a phone call.</p> <p>We needed sub-second delivery. Here&rsquo;s what we built.</p> <h2 id="why-polling-is-the-wrong-model-for-alerts">Why polling is the wrong model for alerts</h2> <p>Polling feels simple: fetch data every N seconds, diff it, surface changes. The problems compound quickly in practice.</p>