Optionally pre-warm the crawl list from Sonarr/Radarr monitored series #6

Description

Re-scoped 2026-07-28 by DR-011.
This was "pull the monitored series list from Sonarr/Radarr — that is the crawl work-list". It is now an
optional pre-warm, because the work-list framing was wrong.

Checking how the comparable projects solve this: neither MediathekArr nor RundfunkArr reads Sonarr's
monitored series via the Sonarr API. MediathekArr is a live search proxy over MediathekViewWeb;
RundfunkArr uses curated shows.json / rulesets.json. Both are pure Newznab providers, because the
Newznab contract is pull-by-query: Sonarr already knows what it monitors and asks for exactly what it
wants via t=tvsearch&q=…&tvdbid=…&season=&ep=.

Two consequences:

• Reach-back is not required for the integration to work.
• Reach-back cannot satisfy search on its own — an interactive search for a show that is unmonitored,
  or monitored but not yet crawled, returns nothing. Query-driven resolution is needed regardless (new
  issue), which makes this an optimisation by definition.

What it is still genuinely good for: pre-warming the standing crawl list so the RSS feed has recent
episodes of the shows the operator actually cares about, without hand-curating a list the way RundfunkArr
does. That is a real UX win over the comparables — just not a prerequisite.

Scope

• Extend IArrClient (added in Sonarr/Radarr instance config UI (Blazor) #4) with a monitored-series fetch: Sonarr GET /api/v3/series
  filtered to monitored: true; Radarr GET /api/v3/movie equivalently.
• Merge the result into the standing crawl list rather than replacing it — a configured
  Crawl:Targets entry must not disappear because an *arr instance went offline mid-poll.
• Map an *arr series to a broadcaster show. This is the hard part, not the HTTP call: TVDB name
  vs. Mediathek topic name, german titles, aliases. RundfunkArr solves it with per-show curated
  rulesets, MediathekArr with a filter/matching system. Expect the same investment here.
• Degrade quietly: an unreachable or unauthorised instance must log and carry on, never fail a crawl
  cycle or empty the list.
• Nothing may depend on an instance being configured (DR-011).

Usage Example

{
  "Crawl": {
    "PreWarmFromArrInstances": true,   // opt-in; default false
    "Targets": [                        // still honoured, and never overwritten
      { "ProviderKey": "ard", "ShowQuery": "Extra 3" }
    ]
  }
}

Alternative

Curate the crawl list by hand — which is exactly what RundfunkArr does and what Crawl:Targets already
supports. Perfectly workable; this issue only removes the manual step.

Notes

• Depends on Sonarr/Radarr instance config UI (Blazor) #4 (instance model + IArrClient) — merged/in review.
• No longer blocks Merged RSS feed for monitored shows #12. Per DR-011 the RSS feed serves the standing crawl list, so Merged RSS feed for monitored shows #12 can proceed
  without this.
• Related: the new query-driven-search issue, which is the part that actually makes t=tvsearch work for
  arbitrary shows.