#41Very Hard
Design a Distributed Tracing System
Asked by:UberAmazon
Recommended Solutions from the Web
Leetcode Discussion
https://leetcode.com/discuss/interview-question/system-design/313117/Design-a-system-capable-of-tracking-bad-request-down-quickly
Great intro to Tracing
https://betterstack.com/community/guides/observability/distributed-tracing/
Dapper research paper
https://research.google/pubs/pub36356/
Solution from FAANG engineer
https://www.youtube.com/watch?v=Jkhw_FSikX0
Key Details & Edge Cases You Should Know
•
How will you collect information from various parts of your backend?
•
Where will you send the info and how will you store it?
•
How will you map a trace for a particular request? Will you use a request id, or something else?
•
If an error happens, how will you surface relevant information to the system admin?
•
How will you handle machines going down?