Effective App Monitoring in Production
summary
Top-down monitoring tracks user actions from viewmodels into a system task object and pushes latency, availability and error metrics to New Relic; dashboards show per-module KPIs and a query revealed a deals-board latency spike where users with >2s loads jumped ~95%, prompting backend parallelization and caching fixes. A second incident finds custom-objects loads repeatedly abandoned due to “cannot access database on main thread” fixed with a dispatcher change. Team responsibility rotates weekly via metrics DRI to review, escalate and iterate.
AI-written summary · about
chapters (20)
- 0:00App monitoring intro
- 2:11Why monitoring matters for user experience
- 4:39Observability strategy
- 6:24Tracking user functions
- 9:36Top-down vs bottom-up tracking
- 11:21KPIs for app reliability
- 13:52Metrics across app layers
- 16:00Tracking in ViewModel
- 18:51Monitoring indicators
- 21:20Metrics with task modules
- 22:43Feature dashboards
- 24:30Monitoring endpoints & exceptions
- 25:53Monitoring process & visibility
- 26:36Metrics DRI role
- 29:25Improving monitoring & debugging
- 30:08Latency case study
- 33:01Untested conditions case study
- 35:33Monitoring recap
- 36:16Team effort in monitoring
- 37:21Q&A
related talks
58mObserving Applications in Production
30mHow to collect application metrics if it needs to be done yesterday
30mRails Performance monitoring: A primer for developers
09mMonitoring? Not OPS problem
52mHolistic Monitoring
32mWhat’s my App *Really* Doing in Production?
52mNot on Our Watch: an Introduction to Application Performance Monitoring
34mMonitoring for devs. The necessity or self-indulgence