Design a Metrics and Alerting System
A metrics and alerting system collects numbers from every service and host, keeps them as time series, draws them on dashboards and pages a person when one crosses a line. Request rate, error rate and p99 latency (how long the slowest 1% of requests take) are the usual examples. The engineer on call gets a phone call at 3 a.m. because a rule said "error rate above 2% for five minutes" and it was true. Microsoft asks this under a few names. "Design an alert monitoring system" shows up in SDE II loops, including one on a Windows team. "Design a distributed monitoring system" and "design a distributed metrics aggregator which can be used by multiple services" are the same system seen from the ingestion side. It also comes up in the LLD round (low-level design, where you write classes and code) as "design an alert system". The design below covers the whole pipeline, from the counter inside a service to the phone that rings.
MicrosoftPremium lesson · System design round
The Microsoft design, worked end to end.
- The API and data modelSystem design · Signature problem
- The design, drawn out2 diagrams
- 6 deep divesWith the tradeoffs
$49 a year · all 107 companies · instant access
1,382 Microsoft interview reports · guide updated
Microsoft interview guideKnow what Microsoft asks before you walk in.
21 coding · 9 system design · 24 behavioral questions
Get their questions first.
InterviewLegend is independent and not affiliated with or endorsed by any company it covers.