Log Collector
logger is an observability tool. It does not own player, Scene, Instance, or routing authority. See ADR-013 in the current architecture decisions.
Current functionality
Applications write local structured logs. APOLLO_LOG_PUSH_ADDR enables existing M1 log push to a collector. logger provides scan/follow queries and a serve collection entry point, with process, severity, time, and category filters.
Collector failure uses bounded buffering and degradation, retaining local logs without blocking simulation. World uses the runtime/log identity world and executable world.
Implementation
- apps/logger: collection and query entry points.
- modules/net/logpush: application log push.
- modules/core/log: structured records and file output.
- tests/test_logger_retrieval.cpp and tests/test_log_push.cpp: existing verification entry points.
Kafka, ClickHouse, and Prometheus/Grafana infrastructure remain requirement-triggered work in the root todo.md. See logging design for storage, query, and failure rules.
