Failing Gracefully As A Feature Lorne Kligerman Director of Product, Gremlin @lklig
2
3
T-Ho 2017 Hey team… bit of a spill but I’m fine. Be down in 10! 4
We Expect Technology To Just Work™ 5
Black Friday Failures Technical Issues Likely Cost Retailers Billions 12.01.16 Macy’s, Lowe’s hit by Black Friday technical glitches 11.27.17 Retail outages online leave shoppers frustrated on Black Friday 11.23.18 @lklig People.com
Breaking Banks Wells Fargo accidentally foreclosed hundreds of homeowners 8.7.18 Customers report difficulty accessing Chase Bank mobile and online 2.16.19 Citibank Website down, not working 2.28.19 @lklig Investopedia
Airline Incidents Computer Problems Blamed For Flight Delays 4.1.19 Major US Airlines hit by delays after glitch at vendor 4.1.19 Pilots of doomed Boeing 737 MAX fought the plane’s software and lost 4.4.19 @lklig
Technology is fragile. When it breaks, we shouldn’t notice. @lklig 9
FAILURE Plan ahead to keep your users happy GRACEFUL DEGRADATION @lklig 10
Why Are Failures So Common? 11
Legacy Systems @lklig 12
Failure UI Lack of Testing End to end Integration Unit @lklig 13
With Scale Comes Complexity @lklig 14
@lklig
@lklig
What Can We Do About It? 17
Design For Failure 18
19 @lklig
Designing For Failure Key User Stories Edge Cases From Dependency Failures & Features Unexpected User Behaviour @lklig 20
@lklig 21 21
Loading Screens Are Not Graceful @lklig 22
Inject Failure By Breaking Things On Purpose @lklig 23
Inject failure one service at a time. Maintain critical functionality. 24 @lklig
Common Failures Modes THAT DEGRADE THE USER EXPERIENCE Errors Blackhole Latency HTTP 400, 401, 402 500, 503 25 @lklig @lklig
Degrade Gracefully 26
Graceful Degradation Provide the best possible experience ● All but the most critical ● functionality can fall off Don’t give up on your users, hold ● state as long as possible 27 @lklig
Auth User Data When one Content dependency fails, users are Storage often affected Cache Feature 1 Feature 2 @lklig 28
Implemented As Designed @lklig 29
Added Latency @lklig 30
Blocked Video Link @lklig 31
Blocked JQuery Request @lklig 32
@lklig 33
Delight Your Users 34
Graceful Degradation Done Right @lklig 35 35
Positive Business Impact Product Launch Success Metrics Product Landing Delight users with Quantitative goals of Were the goals achieved? new features the launch Why or why not? What got in way? @lklig 36
Plan Experiments Early Maintain release Engineers spend less Deliver a positive velocity time in war rooms user experience @lklig @lklig 37
Graceful Degradation As a Feature RELIABILITY THROUGH CHAOS ENGINEERING Design for Failure Inject Failure Degrade Gracefully Delight Your Users Identify the most Impact your system to be Plan for non critical Your product metrics will critical end user sure your user experience functionality not to show behaviour, no functionality. isn’t impacted. get in the way. matter the condition. @lklig 38
USE LORNE FOR 20% OFF
gremlin.com/lorne 40
Q&A Lorne Kligerman Director of Product, Gremlin @lklig
Recommend
More recommend