AppSphere 15 - Achieving Stability and End-to-End Monitoring
0 likes647 views
The document details a presentation by Stace Baal from eHarmony Inc. on the implementation of AppDynamics to improve system stability and monitoring. It addresses challenges such as data fragmentation, visibility into legacy applications, and the need for proactive alerts to prevent outages. The rollout process included defining business transactions and creating dashboards for comprehensive monitoring, aiming to enhance the overall user experience and optimize engineering efforts.
AppSphere 15 - Achieving Stability and End-to-End Monitoring
- 1. Achieving stability and End to End Monitoring Stace Baal, eHarmony Inc. Director of Engineering
- 2. The Problem The Mess or Miss of Data What we’re working on The Rollout The Business Transaction The Drill Down Copyright © 2015 AppDynamics. All rights reserved. 2 Dashboards Health Rules Proactive That shouldn’t happen?! Out of the datacenter – EUM What’s next?
- 3. THE PROBLEM
- 4. The Problem
- 5. The Problem Mature, large brand in the midst of changing architecture • Instability from Monolithic legacy application • Reactive not proactive with production issues • Little visibility into how new services interacted with legacy apps • Lack of data on customer experience
- 6. The Problem
- 7. THE MESS OR MISS OF DATA
- 9. The mess of data We needed one source to truth to quickly diagnose issues • Fragmentation of monitoring systems • Lots of holes, lack of time for maintenance as services evolve • Ineffective alerting, noise • User facing outages
- 10. The mess of data Copyright © 2015 AppDynamics. All rights reserved. 10
- 11. Why APM? Challenges that caused us to choose AppDynamics • Stability • Time to Market: Needed a turnkey solution • Minimal internal resources • Engineering driven solution • Ability to track and trend baselines • Single pane of glass
- 12. THE ROLLOUT
- 13. How do we roll this out? Challenges that caused us to choose AppDynamics • Deploy • Define • Diagnose • Notify • Discovery Tiers Business Transactions Dashboards Health Rules Snapshots
- 14. Rolling out AppDynamics Decide what to cover and how to divide it up • Create applications • Define and organize tiers • Deploy – Start Front to Back – Chef Integration – Staging – Production, rolling deploy • Wow, that went fast… – Time to market vs. homgrown solutions
- 16. Rolling out AppDynamics Copyright © 2015 AppDynamics. All rights reserved. 16 • 4 Applications – User Front – User Back – Batch – Staging • 38 Tiers (and counting) • EUM, Web & Mobile
- 18. Defining Business Transactions Copyright © 2015 AppDynamics. All rights reserved. 18
- 19. Defining Business Transactions BT: Business transaction correlated across service tiers • Visibility into what is most important • Start simple, broad transactions key to the business • Add granularity over time • May take some time, but it’s worth it
- 20. Defining Business Transactions Copyright © 2015 AppDynamics. All rights reserved. 20 One Transaction Several URLs
- 21. Defining Business Transactions Copyright © 2015 AppDynamics. All rights reserved. 21 One Transaction Correlated Tiers
- 22. Defining Business Transactions BT is the foundation for all diagnostic tools • Correlation • Trends and dynamic baselines • Health rules • Drilldown and discovery
- 23. Defining Business Transactions Copyright © 2015 AppDynamics. All rights reserved. 23
- 24. THE DRILLDOWN
- 25. Drilling down into Transaction Snapshots Your diagnostic window into a transaction • Single pane view across tiers • Deep code-level visibility • Database and exit calls • Custom diagnostic data • Server health and utilization
- 26. Drilling down into Transaction Snapshots Copyright © 2015 AppDynamics. All rights reserved. 26
- 27. Drilling down into Transaction Snapshots Copyright © 2015 AppDynamics. All rights reserved. 27
- 28. DASHBOARDS
- 29. Dashboards
- 30. Dashboards Your tool to bring everything together • At a glance diagnostic and status • Comprehensive tier and node monitoring • Connecting the dots between systems • Trending over time • Easily surface data between teams
- 31. Dashboards Copyright © 2015 AppDynamics. All rights reserved. 31 Traffic Performance
- 32. Dashboards Copyright © 2015 AppDynamics. All rights reserved. 32 WEB API Service
- 33. Dashboards Copyright © 2015 AppDynamics. All rights reserved. 33
- 34. PROACTIVE
- 35. So now what? • Data and tools drive a cultural shift • Alert and surface issues before than can become outages • Health Rules are able to be predictive • Socialize the system beyond on-call • Empower teams and increase visibility • Encourage engineers to drilldown and optimize
- 38. That shouldn’t happen?! That Random Spike Copyright © 2015 AppDynamics. All rights reserved. 38
- 39. That shouldn’t happen?! After the release Copyright © 2015 AppDynamics. All rights reserved. 39
- 40. OUT OF THE DATACENTER - EUM
- 41. End user Monitoring Extending visibility to the end user experience • Web: Pages & Ajax Requests • Mobile: Network Requests • Mobile: Crash Reporting • Custom Events & Timers • User snapshots correlated with BT snapshots • Justify and prioritize infrastructure and application enhancements • Correlate performance to business metrics • CDN and 3rd party monitoring
- 42. Stability & End to End Monitoring Copyright © 2015 AppDynamics. All rights reserved. 42
- 43. WHAT’S NEXT?
- 44. Thank You