Design a metrics platform without alerting
Company: LinkedIn
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Onsite
Design a **metrics platform** (like a simplified Prometheus/Datadog metrics product) that supports collecting, storing, querying, and visualizing metrics.
### Scope
- Ingest time-series metrics from many hosts/services.
- Store metrics with tags/labels (dimensions).
- Support queries for dashboards (instant queries and time-range queries).
- **No alerting/notification system is required.**
### Clarifications to cover
- Data model (counters, gauges, histograms; labels/tags)
- Ingestion model (push vs pull)
- Storage (hot vs cold, retention, downsampling)
- Query API and typical query patterns
- Scale assumptions and bottlenecks (write QPS, cardinality)
- Multi-tenancy, authn/authz, and isolation
- Reliability and cost controls
Quick Answer: This question evaluates a candidate's ability to design a scalable time-series metrics platform, covering data modeling for counters/gauges/histograms and labels, ingestion and storage strategies, query APIs, multi-tenancy, and operational trade-offs like reliability and cost.