Problem
Discovery (POST /api/discovery/scan) is on-demand only — it has to be triggered manually per CIDR from the Discovery workspace. For keeping a homelab inventory fresh, I'd like rackpad to re-scan a set of subnets automatically on a schedule, without an external cron driving the API.
This is especially useful in environments where L2/ARP MAC capture isn't available (e.g. Kubernetes pods on a routed CNI, Docker bridge, routed VLANs). The L3 ICMP ping-sweep works fine there, so a periodic L3 sweep is a great lightweight alternative to ARP/MAC discovery for tracking what's alive.
Proposed solution
- A configurable list of CIDRs to scan periodically (per-lab), with an interval — conceptually similar to the existing
MONITOR_INTERVAL_MS monitoring loop, but for discovery instead of known monitors.
- Reuse the existing
/scan logic (ICMP sweep + optional MAC enrichment where available), writing into the same review-first discovered-devices staging list.
- Per-CIDR enable/disable + interval, mirroring how device monitors already support per-monitor
intervalMs.
Notes / questions
- Today the scan is capped at
/24 or smaller per call — scheduled scans of larger ranges would presumably fan out into /24 chunks internally.
- Results would stay review-first (no auto-import), which is fine — the goal is just keeping the discovered list current automatically.
- An optional long-lived API token (instead of session-only
Bearer) would also make external scheduling easy as an interim workaround, if a built-in scheduler is out of scope.
Problem
Discovery (
POST /api/discovery/scan) is on-demand only — it has to be triggered manually per CIDR from the Discovery workspace. For keeping a homelab inventory fresh, I'd like rackpad to re-scan a set of subnets automatically on a schedule, without an external cron driving the API.This is especially useful in environments where L2/ARP MAC capture isn't available (e.g. Kubernetes pods on a routed CNI, Docker bridge, routed VLANs). The L3 ICMP ping-sweep works fine there, so a periodic L3 sweep is a great lightweight alternative to ARP/MAC discovery for tracking what's alive.
Proposed solution
MONITOR_INTERVAL_MSmonitoring loop, but for discovery instead of known monitors./scanlogic (ICMP sweep + optional MAC enrichment where available), writing into the same review-first discovered-devices staging list.intervalMs.Notes / questions
/24or smaller per call — scheduled scans of larger ranges would presumably fan out into/24chunks internally.Bearer) would also make external scheduling easy as an interim workaround, if a built-in scheduler is out of scope.