Design a Logging and Monitoring Platform
Ingest high-volume telemetry reliably, separate hot search from cheap retention, and make alerting useful rather than noisy.
Telemetry platform: agents batch into an ingestion layer, durable streams feed hot search, metrics and cold archival
Agents batch and compress logs, metrics and traces to an authenticated ingestion gateway. A durable stream absorbs bursts; processors index a bounded hot window and archive older data cheaply. Partition by tenant/time, enforce quotas and sample high-cardinality telemetry.
Alerts should use user-facing SLO symptoms and link to runbooks. Preserve tenant isolation, retention controls and an audit trail. Measure ingestion lag, dropped samples, query latency, cardinality and storage cost.