Editor's note
Modern, distributed applications feature significantly more complex component interactions than the legacy monoliths with which IT organizations are familiar, and virtualized infrastructures, such as cloud platforms, operate differently than physical servers. Organizations must rethink their application performance improvement methods.
The application performance management (APM) umbrella encompasses monitoring, error resolution and multiple IT and non-IT departmental responsibilities. Start with an evaluation of available APM tools, based on what your organization needs and its budget, and take into account monitoring focuses, metrics and alert preferences. Proceed with a close interrogation of your organization's processes and methodology -- who is involved and who should be.
1Fine-tune processes and methodology
Application performance improvement goes beyond simply having the right tool to capture ops data. Learn how to direct attention to the most likely sources for a bottleneck, which team members to involve in the process and how to ensure that cloud applications float along smoothly.
-
Article
Recalibrate the focus of app performance metrics
IT teams build, organize and operate applications and the resources they live on differently than the last generation of apps. To meet user expectations and demands on these new architectures, transition from traditional key performance indicators to a more UX-focused lens. Read Now
-
Article
Tweak hardware configs and resource allocations
When applications don't run as well as expected, look at resource allocation and host health, along with other factors. After any change, monitor subsequent activity closely, and compare to historical data to guide the way to a better application. Read Now
-
Article
Where's Waldo -- or, rather, the performance bottleneck?
Finding the source of a performance slowdown can be a frustrating Where's Waldo puzzle. Use these pointers on the most likely hiding spots to save some headaches. Read Now
-
Article
You're all in this together
Rely on application owners to fill in knowledge gaps, and empower internal superusers to work alongside the help desk team. People with technical knowledge outside the IT team can answer peers' questions and offer insights that facilitate application performance improvement. Read Now
-
Article
Cloud application performance midterms
When it comes to exams, you're better off if you study all the possible questions. The same goes for applications in the cloud. Test applications for performance under load and stress, as well as these three other key factors. Read Now
-
Article
Continuous delivery runs more smoothly with APM
While it can seem like CD and APM are at odds, the process for application performance improvement can scoot leftward in the prerelease stage of the CD pipeline. It's easier to fix a problem that hasn't affected users than a problem that's already live. Read Now
2Comprehensive monitoring is key for APM
An organization committed to application performance improvement must be equally committed to a monitoring setup that tracks the right metrics and sends out the right alerts. It's unlikely that any tool can do these things perfectly off the shelf; these tips aid in customization and tuning.
-
Article
Draw the baseline
Monitoring tools use either static or dynamic thresholds to notify IT when there's a metric out of spec. Both methods have pros and cons. Read Now
-
Article
Distinguish a fault from a bottleneck
Admins must recognize the difference between faults and bottlenecks. Ensure monitoring data is precise and comprehensive, and be prepared to comb through layers of application and host data to find the source of an issue. Read Now
-
Article
Smart log analysis yields stronger apps
AI-fueled log analytics and monitoring tools offer organizations a range of ways to evaluate and interpret data and boast other benefits not found in basic tools. Let the algorithms point the way to application performance improvement. Read Now
-
Article
SNMP monitoring broadens IT's view
SNMP can track both infrastructure and application metrics at once. And because SNMP is open source and widely used, most major vendors provide the necessary tools to adopt this method. Read Now
-
Article
Distributed apps: Attempt to catch them all
Distributed applications make IT monitoring difficult -- it's like trying to count chickens running around in the yard. Not only do microservices shift, they often disappear when inactive. Predictive analytics correlated to UX can't let IT see everything all at once but can suggest what to examine. Read Now