One of the worst things to experience as a system administrator is to get that alert in the middle of a Friday or Saturday night. Without even taking into consideration ever-encroaching problems of alert fatigue, emergency alerts will typically reflect a front-end problem users are experiencing. Subpar user experience, of course, also very likely portends lost revenue as users understandably take their online experience elsewhere.
In “Strategies of Top Performing Organizations in Deploying AIOps,” a study by analyst firm Digital Enterprise Journal (DEJ) based on the results of over DevOps professionals from over 1,100 organizations, over 90% of the respondents reported lost revenues due to slow speeds in application performance. Also reported in the survey was an average monthly loss of $634,000 due to slowdowns in application performance and $1.27 million lost annually devoted to managing avoidable incidents.
The main takeaway is that a monitoring platform must also issue performance alerts earlier before front-end performance becomes a real issue, as well as communicate to developers what must be done to fix the code.
While infrastructure monitoring remains important for operations-related performance of the network, servers and applications, such as database and storage performance, developers generally have very little information into the direct effect their code has on front-end performance.
“One thing developers will tell you is that they don’t have visibility into how their code is doing in production,” Milin Desai, CEO for Sentry, told The New Stack. “You push out code multiple times a day” without direct information about how a user might have difficulties performing a transaction or an application is not loading fast enough.
“Essentially, in real-time, we tell the developer or the development team that these many users are hitting this issue on iOS version 13 and it’s primarily happening to users in these countries,” Desai said.
“We were primarily focused on finding software errors — now we are going to help you understand the performance of your ports,” Desai said. With the new release, “a developer can now say ‘this last API call is taking 10 seconds to complete, so maybe we want to look at it.’ The infusion is now being converted into this visibility of what to fix consisting only four lines or five lines of code.”
Specific performance-monitoring insights the release offers that Sentry communicated include:
- Application Health Insights: application response time to their interactions with live updating latency and throughput data and comparisons of slow response times, increases in transactions and error rates to diagnose and fix performance issues.
- Transaction Summary View: transactions are sorted by slowest duration time, related issues, and the number of users having a slow experience in a consolidated view.
- Root Cause Analysis: identifications of differences in characteristics between outliers and normal performing transactions with drill-down capabilities and user-friendly visualizations.
- Tracing Leverage: distributed tracing to reveal the database query that caused an error or performance issue.
- Performance Alerts: See how crashes contribute to performance and set thresholds to get alerted if performance metrics fall past a predefined tolerance band. Drill down into transaction details within tracing waterfalls, which visually highlight API call times in relation to expected operations and device data, to quickly identify which API calls are giving customers poor experiences.
Sentry is a sponsor of The New Stack.
At this time, The New Stack does not allow comments directly on this website. We invite all readers who wish to discuss a story to visit us on Twitter or Facebook. We also welcome your news tips and feedback via email: firstname.lastname@example.org.