New performance monitoring tool alerts developers in real time

Application monitoring tools provider Sentry has introduced a new performance monitoring product to help developers find and fix performance issues by adding just five lines of additional code to their application.

San Francisco-based Sentry provides tools that development teams use to identify performance issues by tracing them to poor performing API calls along with related errors.

The tool allows developers to understand how their code is doing in either production, preproduction or in staging, said Milin Desai, CEO of the company. "Essentially, we help developers find errors in their software in real time," he said.

Performance bottlenecks and code errors can mean lost business for organizations when customers can't access services. Indeed, a recent study published by researchers at Digital Enterprise Journal shows that 91% of organizations have reported missed revenue due to performance and availability issues and the average monthly loss is an estimated $634,000.

With the Sentry technology, when an error occurs the system will notify the development team using Slack, PagerDuty, email or whatever the notification system of choice is. The notification includes how many users are affected, what devices and browsers are affected and all the metadata about the issue. In addition, the notification is broken down to the point of identifying the exact line of code and the potential code comment where that issue could have come from.

This enables developers to quickly fix the code and roll it back out. Sentry is available as a SaaS service. There are more than 60,000 development teams using it, Desai said.

The new Performance Monitoring feature builds onto Sentry's error tracking and release integration platform to help developers resolve performance bottlenecks quickly. The performance monitoring feature is included in the Team, Business and Enterprise versions of Sentry.

"Automated code performance monitoring is critical in today's agile world," said Krishnan Subramanian, an analyst at Rishidot Research in Redmond, Wash. "If the goal of Agile development and DevOps is to take the applications faster to market, automated error detection and performance of the code are important."

There are alternatives to Sentry with varying capabilities, such as OverOps, Honeybadger and Rollbar.

Charlotte Dunlap

These application performance monitoring (APM) capabilities could show up in more app management tools as enterprises seek tools to support DevOps, said Charlotte Dunlap, a GlobalData analyst based in Santa Cruz, Calif.

"IT operations teams are increasingly seeking APM tools as part of their app modernization efforts, looking to improve app performance through monitoring tools aimed at a broader audience, which includes developers," she said. "As such, we'll begin to see more consolidation of this type of APM technology with larger application management solutions as enterprises seek tools that support more of a DevOps model."

The new offering provides Application Health Insights that show users how their app measures up with live updating latency and throughput data, such as increases in transactions and error rates. It also offers a transaction summary that details transactions sorted by factors such as slowest duration time and the number of users having a slow experience.

"I wouldn't say this is groundbreaking, but it boosts productivity significantly and helps developers detect errors before an app even gets into production," Subramanian said. "Most people think of Agile as taking the code faster to production, but, for me, Agile is about taking good code faster to market. From this point of view, tools like Sentry add significant value to any organization."

Moreover, Sentry's latest improvements are more focused on app performance and user experience, but there are security implications a platform like Sentry can deliver as well.

The top 10 web app security risks are largely the same today as they were during the second Bush administration -- injection flaws, busted authentication and access controls, and cross-site scripting, said Chris Gonsalves, senior vice president of research at The 2112 Group in Port Washington, N.Y.

Organizations don't necessarily need super-fancy AI-driven 'wonder tools' to unearth this stuff. They need basic blocking and tackling, which Sentry handles very well.

"It's the same old stuff," Gonsalves said. "Organizations don't necessarily need super-fancy AI-driven 'wonder tools' to unearth this stuff. They need basic blocking and tackling, which Sentry handles very well.

"Add to this Sentry's ability to ingest, aggregate and report Content-Security-Policy violations, Expect-CT and HTTP Public Key Pinning failures, and what you have is a valuable set of capabilities to improve basic app hygiene, something most development organizations could definitely use."

The product's Root Cause Analysis feature helps users to see the differences between outliers and normal performing transactions. It also provides Tracing and Performance Alerts that check for performance metrics that fall below certain predefined thresholds.

Sentry's core error monitoring product covers all popular programming languages. The new performance monitoring tool supports just JavaScript and Python, and Sentry will be adding additional languages every quarter, Desai said.

The base plan for Sentry's software starts at $26 per month and includes 50K errors, 100K transactions, 1 GB attachments for a team, then works its way up depending on if it's for a team or business, number of errors and other factors.