Engineering Reliable Mobile Applications
summary
Argues that mobile SRE differs from server SRE and gives concrete failure modes: a doodle config missing color causes crashes across time zones and caches keep bad config alive; a tiny legacy config change breaks many old clients forcing manual upgrade; a Play Services update triggers repeated Firebase registration spikes that exhaust quotas. Recommends measurable practices: staged rollouts, experiments and feature flags, monitor client-side SLIs and correlate by experiment ID, avoid upgrade‑proportional load and add server guardrails.
AI-written summary · about
chapters (23)
- 0:00Welcome
- 1:17Introduction
- 2:23Agenda and SRE basics
- 3:06Client-side reliability matters
- 5:09Mobile SRE challenges
- 6:15Monitoring mobile apps
- 7:19Control limits in mobile
- 8:00Mobile app availability
- 9:23Telemetry and crash reports
- 10:27Real-time metrics delays
- 11:30Measuring production behavior
- 12:33Resource efficiency and Android vitals
- 13:36Change management importance
- 14:38Staged rollouts and beta
- 15:19Experimentation and A/B
- 16:03Feature flags for releases
- 17:29User upgrade behavior
- 18:32Client-server risks
- 19:15Case study: Doodle config crash
- 22:05Case study: Legacy config outage
- 25:34Case study: Upgrade load
- 28:21Key takeaways
- 29:25Actionable recommendations
related talks
04mThe SRE I aspire to be
17m“How’s the App Doing?” Bringing Mobile Into Your Reliability Picture
18mProduct Reliability in
47mSRE Team Lifecycles
40mWhy Is My App SLOw? Defining Reliability in Platform Engineering
35mSRE 101
32mThe culture of Site Reliability Engineering
25mElevating Android App Resilience: Incident Management Strategies