-
Notifications
You must be signed in to change notification settings - Fork 0
FAQ and Troubleshooting
No. Nudgarr sits alongside them and tells them what to search via their existing APIs. It does not manage your library, rename files, or handle downloads — that is still entirely handled by your arr apps and download client.
No. Each sweep nudges up to the configured Max per Run cap per pipeline, and items searched recently are skipped due to the cooldown. By default, each run nudges at most 1 Cutoff Unmet movie and 1 Cutoff Unmet episode per instance, with a 48-hour cooldown preventing the same item from being re-searched within that window.
Make sure Automatic Sweeps is toggled on in Settings. By default it is off. Once enabled, the first sweep fires at the next cron trigger, not immediately.
History writes happen during the sweep. If nothing appears, it is likely that all eligible items were filtered out — either on cooldown, already queued, excluded, or the Cutoff Unmet pipeline is toggled off. Check the Sweep panel stats for the relevant counts.
Nudgarr could not reach the instance during the last sweep. Common causes:
- The URL is wrong or contains a trailing path
- The API key is incorrect or has been rotated
- The arr app is down or unreachable from the Nudgarr container
- A network or firewall rule is blocking the connection
The URL should be the base URL only — e.g. http://192.168.1.10:7878 with no trailing slash or path.
Overrides only take effect once you Apply them on each instance card in the Overrides panel.
Disabled instances are skipped in sweeps. Check the toggle state in the Instances panel and save again if needed.
The defaults are intentionally conservative. Increase Max per Run in the Settings panel to search more per sweep.
Nudgarr can only trigger the search — whether a result is found and grabbed depends on your indexers, quality profiles, and download client. Check the arr app's search logs for the relevant titles.
Verify that Cooldown (Hours) is set to a non-zero value in the Settings panel. The default is 48 hours.
Check the Cutoff Unmet toggles in the Settings panel. There are separate toggles for Radarr and Sonarr. If either is off, that app's Cutoff Unmet pipeline is disabled entirely regardless of Max per Run.
The import check runs on its own independent timer (default: every 120 minutes). It may not have fired yet. You can trigger an immediate check from the Library panel Imports view.
This means the same item has been imported more than once — an upgrade from a lower quality to a higher quality after the initial grab. Hover over the number to see the full quality upgrade path.
The most common cause is that cutoffFormatScore is set to 0 (or not configured) on your Sonarr quality profiles. If there is no score cutoff defined, every episode passes the gap check and nothing is indexed. Open Sonarr, go to Settings > Profiles, and set a non-zero custom format score cutoff on the profiles you want covered.
Another common cause is that all your series have a tag that matches an excluded tag in the Nudgarr Filters panel. Check your Sonarr instance filters — if a tag applied to all your series is in the excluded list, every series will be filtered before the sync can index anything.
Yes — back up /config/nudgarr.db (SQLite database) and /config/nudgarr-config.json. These two files contain everything.
If the container cannot parse the config file on startup, it falls back to defaults. Check the container logs for validation errors. The original file is left in place — you can inspect and correct it manually.
Deleting an instance removes it from sweeps and health checks but does not remove its historical data. Use Clear History and Clear Imports in the Advanced panel Danger Zone for a clean slate.
Expected on a fresh install or after Reset Intel. Intel waits until 25 confirmed imports or 50 completed sweep runs have accumulated before displaying data. All cards appear once either threshold is met.
This is expected. Reset Intel clears the lifetime aggregate back to zero, which puts Intel below the cold start threshold again. Data begins accumulating immediately on the next sweep and import check cycle. Depending on how active your install is, it may take a few sweeps before the threshold is met again.
Intel data is stored in a protected aggregate that is not touched by Clear History or Clear Imports. If the numbers look unexpected after a clear, the Intel aggregate still holds the accumulated data from before the clear. This is correct behaviour.
For a fully clean slate, use Reset Intel (Intel panel) after clearing history and imports.
The CF Score Health card is only visible when CF Score is enabled. Enable it in the Pipelines panel to unlock it.
Auto this month counts auto-exclusions since the first of the current calendar month in UTC. If your local timezone is significantly offset from UTC the boundary may appear to shift by several hours. The count resets naturally at the start of each new UTC month without deleting any data.
This is normal and expected when CF Score is your most active pipeline. CF Score searches files that already exist in your library -- Radarr or Sonarr finds a better-scored release very quickly once told to search, often within minutes. The fast turnaround is the pipeline working correctly, not a display error.
For example, Bluray-1080p to Bluray-1080p. This is CF Score working correctly. The quality tier did not change but the custom format score improved -- Nudgarr found a better-scored release within the same quality tier. The upgrade path tracks the quality profile tier labels, not the CF score values, so same-tier CF Score upgrades appear as matching labels on both sides.
Check that at least one day of the week is selected and that the TZ environment variable is set correctly. The hint line below the time inputs in Settings will confirm the window is being interpreted as expected.
No. Run Now always fires immediately regardless of the maintenance window.
This is expected. The maintenance window suppresses scheduled sweeps and CF Score syncs only. Import checking continues during the window because it only contacts your local Radarr and Sonarr instances, not indexers. If something started downloading before the window and finishes during it, the import check will still catch it and confirm it promptly.
Auto-exclusion runs inside the import check loop, not during the sweep. A title will not be evaluated until the import check interval elapses after the sweep that pushed it over the threshold. To keep auto-exclusion responsive, set your import check interval equal to or shorter than your cron schedule.
This was caused by a bug in versions prior to v4.3.0. The queue check used a per-item API endpoint that some Radarr and Sonarr versions silently ignore, causing every candidate to appear as actively downloading and be skipped indefinitely. The fix in v4.3.0 fetches the full queue once per instance instead. Affected titles will be correctly evaluated on the next import check cycle after upgrading.
This is a timing edge case — the title was grabbed after the last import check but before the current one ran, so the import was not yet confirmed when auto-exclusion evaluated it. To recover: find the title in the Library panel Exclusions view and click Remove. The search count resets to 0 automatically.
In the Library panel Exclusions view, click Clear Exclusions and select Clear Auto. A confirmation dialog appears before the action runs.
Try a hard refresh (Ctrl+Shift+R / Cmd+Shift+R). If still broken, check the container logs for startup errors.