## Context Monitoring currently needs retention settings that can vary by product plan: - Starter: shorter history, e.g. 7 days - Scale: longer history, e.g. 6 months The broader monitoring roadmap already tracks configurable retention in #14. This issue scopes the concrete Docker Compose knobs implemented on branch `claude/check-data-retention-UkR7s`. ## Proposed Change Expose retention as runtime configuration: - `VM_RETENTION_PERIOD` for VictoriaMetrics `-retentionPeriod` - `QUERYID_RETENTION_HOURS` for queryid mapping retention in the monitoring Flask backend Defaults should preserve current behavior: - VictoriaMetrics retention: `336h` - queryid mapping retention: `720` hours ## Acceptance Criteria - Docker Compose accepts `VM_RETENTION_PERIOD` and passes it to VictoriaMetrics. - Docker Compose accepts `QUERYID_RETENTION_HOURS` and passes it to the monitoring Flask backend. - `.env.example` documents both variables, defaults, and suggested Starter/Scale values. - Existing deployments keep the same effective retention when the variables are unset. ## Related - Parent roadmap item: #14 - Implementation branch: `claude/check-data-retention-UkR7s`