Breakpoint recap: Uptime Monitoring, robots, and feature flags galore

Bugs don’t announce themselves politely. They crash your checkout flow, break authentication, or slow your API to a crawl—usually right before your CEO asks how things are going. And when the error inbox is flooded with a hundred variations of TypeError: cannot read property of undefined, figuring out what actually matters can feel impossible.

That’s why we’ve rolled out updates to eliminate the guesswork in debugging—so you can spend less time searching and more time building. Get alerted to downtime before it impacts your users, see related issues and user feedback directly when on your site, cut through duplicate alerts with AI-powered grouping, and resolve issues faster with AI-assisted fixes. 

Catch application outages before your boss does

A production outage is bad. Finding out about it from your CEO’s Slack message? Worse. Uptime Monitoring, now generally available, helps you catch outages before you’re blown up on social (or they make their way to the company all-hands meeting).

With global checks running as frequently as every 60 seconds, you’ll know the second something breaks. You can:

• Monitor critical flows—like checkout pages to detect if purchases suddenly stop processing, or login endpoints to see if users are getting locked out

• Set custom conditions—like alerting only if downtime lasts over 60 seconds

• Get notified where you work—via Slack, PagerDuty, Teams, or webhook

For example, let’s say your checkout page suddenly starts timing out. Instead of waiting for customers to complain, you get an alert immediately. You check the failure logs and see that a third-party payment service is responding slowly. By correlating with Sentry errors and traces, you pinpoint exactly where the slowdown is happening—whether it’s an API failure, a network issue, or something in your own code—so you can fix it before lost sales start adding up.

We’ve already detected 40,000 downtime incidents to date, helping teams catch issues before they turned into a flood of support tickets. 

You get one monitor for free and additional monitors are just $1 each per month. You can create custom alerts for specific URLs with full control over request details, including HTTP method, headers, and body parameters. Try it out or read the blog to learn more.

Stop digging through duplicate errors

Your error inbox is overflowing, but are you actually looking at 10 different issues, or just 50 variations of the same one? Is the alert that just interrupted your flow only a reheat of yesterday’s (error) soup? Instead of wading through an endless stream of alerts, issue grouping automatically clusters related errors, cutting down the noise so you can focus on fixing what matters.

If your API is throwing 500 Internal Server Error across different parts of your app, issue groupingAI Grouping can determine whether they stem from a single root cause or multiple distinct problems. It also groups errors across environments, so you’re not wasting time chasing down the same issue twice in staging and production.

In the largest-ever improvement to our issue grouping algorithm, we recently rolled out a new, ML-powered approach that reduces new issue creation by 40%.  Built on a custom transformer-based text embeddings model, it maintains great performance with sub-100ms processing latency—so you get smarter issue grouping, and less alerts, without slowing down your workflow. 

Triage faster with a redesigned issue details page

You get an alert. Fantastic. Now what? Instead of jumping around to piece together stack traces, user impact, and performance data, the new issue details page puts everything in one place—so you can fix issues faster and with fewer distractions.

When an error occurs, you can now:

• See embedded trace information to diagnose errors and performance issues without switching views.

• View clearer stack traces to pinpoint the root cause faster.

• Check user impact summaries to prioritize issues based on real-world impact.

So now, if a recent deployment slowed down API responses, you can immediately see whether it's affecting 1% or 30% of users—a strong hint of whether it’s just an annoyance or a full-blown emergency. Then, jump straight to the stack trace, tags, and context needed to fix it faster.

To try out the new experience, click the “flask” icon in the upper right of your issue details page to turn it on. Have thoughts? Share feedback directly from the same menu—we're listening.

Understand errors fast. Fix them even faster.

You see AttributeError: 'NoneType' object has no attribute 'group'. Cool. What does that mean? Instead of manually debugging, AI Issue Summaries provide a concise explanation of the error and its likely causes, so you don’t have to piece it together yourself.

For example, if your backend throws KeyError: 'user_id', the summary might explain that:

• This happens when a dictionary lookup fails due to a missing key.

• It’s likely caused by an unexpected API response or missing data.

That’s useful. But knowing what’s wrong doesn’t always mean knowing how to fix it. 

That’s where Autofix comes in.

How it works

Let’s say your app throws TypeError: cannot read property of undefined. Instead of combing through logs, you open the issue in Sentry and Autofix does the hard part for you. It:

• Pinpoints the exact location of the issue in your code.

• Suggests how to handle missing values safely.

• Provides a recommended fix that you can apply directly.

You can review the fix, tweak it if needed, and even create a draft pull request—all from within Sentry. No need to jump between tools or manually test different solutions.

Since launching, Autofix has found the correct root cause 94.5% of the time and successfully fixed the issue with its suggested code changes 53.6% of the time.

Issue Summary and Autofix are currently in Beta, free of charge, for paid Sentry customers. To get started, opt into Sentry AI. You can find it in the Solutions Hub on the issue details page—just click into any error from the Issues page.

Debug in real time from your site—without switching between tools

You just rolled out a new checkout button behind a feature flag. A user reports it’s broken, but when you test it, everything works fine. Instead of digging through logs and trying to reproduce the issue manually, open the Dev Toolbar (in Beta) to see live errors, feature flags, and user feedback for the page in question.

To troubleshoot, you can:

• Check feature flags—was the button actually enabled for them, or are they stuck in an old variant?

• See live errors—did a JavaScript exception prevent the click from triggering?

• Review user feedback—are multiple users reporting the same issue, or is it an isolated case?

Since the Dev Toolbar works in local, staging, and production, you can catch issues early and resolve them faster. Follow the setup instructions here to get started.

Track down issues caused by new feature rollouts

A new feature goes live and suddenly users are reporting performance issues. Is the feature causing it, or is it a coincidence? Feature Flag support in Sentry (in beta) helps you quickly tell whether an error happened in an experiment group, a staged rollout, or an unexpected flag state. 

If you just rolled out a new checkout experience and errors start appearing, you can:

• See which feature flags were active when the error occurred

• Compare error rates before and after the rollout

• Quickly disable the flag if it’s causing issues

Instead of guessing whether the latest deployment is responsible, now you’ll know. To get started with evaluation and change tracking, read the docs here. 

Debug faster, fix smarter

With these updates, Sentry makes it easier to catch issues before they escalate, understand their impact, and fix them fast. While other tools overwhelm you with data, we focus on getting you to resolution as quickly as possible—so you can spend less time troubleshooting and more time shipping.

Head over to your Sentry account to try these updates out, join the conversation on Discord—and help shape where these features go next.

Don’t have a Sentry account yet? No problem—you can try it for free or request a demo to get started.