Design a high-level centralized logging system for a company. Assume many services (microservices, batch jobs) running across multiple hosts/containers need to emit logs. The system should: - Ingest logs from many producers. - Store logs durably with retention. - Support querying/filtering by time range, service, host, log level, request/trace ID. - Handle bursts and partial outages. - Provide basic access control and multi-tenant separation (e.g., by team/environment). Describe the architecture, key components, data model/indexing strategy, scaling approach, and reliability considerations. No diagram is required, but be clear about data flow.