Feature Idea: Per-Instance Queue Threshold #4

Overview

A per-instance queue threshold setting that prevents Nudgarr from triggering searches when a Radarr or Sonarr instance already has a full download queue. If the queue exceeds the threshold, that instance is skipped for that sweep and retried at the next interval.

Why it's useful

• Avoids unnecessary indexer load when downloads are already queued
• Helps on slower connections or high I/O systems where pile-up causes strain
• Keeps Nudgarr from searching for content that won't be acted on immediately

How it would work

• Configured per instance in the instance modal — a single number field defaulting to 0 (disabled)
• Before sweeping an instance, Nudgarr calls /api/v3/queue?pageSize=1 to get the current queue count
• If the count exceeds the threshold, the instance is skipped and a notification fires (optional, follows existing notification pattern)
• The sweep continues normally for all other instances

UI behaviour

Desktop — skipped instance shows an amber "Queue Full" pill in the Sweep tab card, cached stats from last successful run remain visible

Mobile — skipped instance shows an amber status dot when collapsed, and "Queue Full · Last run: Xh ago" text when expanded — consistent with how disabled instances are handled

Implementation scope

• arr_clients.py — two new queue count functions
• sweep.py — threshold check per instance, early return if over threshold
• ui.html — one field in instance modal, one notification toggle, sweep card rendering for both desktop and mobile
• No breaking changes — defaults to 0, existing installs unaffected