fix compression

README.md

@@ -42,16 +42,22 @@ docker run -p 8080:8080 -v $(pwd)/data:/app/data myapp:test
 
 Add services from the dashboard (e.g. `https://example.com`, `google.com:443` for TCP) and view reports.
 
-### Check Retention
+### Check Retention and Rollups
 
-To limit database growth, the app prunes old checks every 15 minutes:
+To limit database growth, the app **rolls up** old checks into hourly aggregates, then prunes raw data:
+
+1. **Rollup** (every 15 min): Checks older than `ROLLUP_AGE_HOURS` are aggregated into hourly buckets (total, success count, latency stats) and stored in `uptime_rollups`. Raw checks in those hours are deleted.
+2. **Prune**: Keeps last `CHECK_RETENTION_COUNT` raw checks per service; optionally deletes by age.
+
+This lets you report accurate uptime over **90+ days** without storing millions of raw checks. Reports combine rollups (historical) + raw checks (recent).
 
 | Env var | Default | Description |
 |---------|---------|-------------|
-| `CHECK_RETENTION_COUNT` | 5000 | Keep last N checks per service |
+| `ROLLUP_AGE_HOURS` | 24 | Aggregate checks older than N hours into hourly buckets |
+| `CHECK_RETENTION_COUNT` | 5000 | Keep last N raw checks per service |
 | `CHECK_RETENTION_DAYS` | 0 (disabled) | Also delete checks older than N days |
 
-Example: keep 2000 checks per service and drop anything older than 30 days:
+Example: keep 2000 raw checks per service and drop anything older than 30 days:
 
 ```bash
 docker run -e CHECK_RETENTION_COUNT=2000 -e CHECK_RETENTION_DAYS=30 ...