*arr stack
I just noticed that the most arr things like sonrar, radarr, lidarr work with a metadata server hosted by the team of the project. For everything we search or want to add this server is connected. Also those servers are behind cloudflare. My first question…
I just noticed that the most arr things like sonrar, radarr, lidarr work with a metadata server hosted by the team of the project. For everything we search or want to add this server is connected. Also those servers are behind cloudflare. My first question is: Why? Why I just found out: I try to import music into lidarr from spotify. It doesn't work so I actived trace log. There I found HTTP/2.0 [POST] https://api.lidarr.audio/api/v0.4/spotify/lookup: 502.BadGateway and wondered. wtf is this? So I curled to it and saw also this: cf-cache-status: HIT report-to: {"group":"cf-nel","max_age":604800,"endpoints":[{"url":"https://a.nel.cloudflare.com/report/v4"}]} So every stupid thing I do is tracked somewhere else, out of my control, no option to opt out. I don't understand why there needs to be a spof added that is out of my control (for what do I selfhost then?) and also disables me to fix a problem by myself. Movies, Series and also songs, can directly requested from other sources without going somewhere else. And this is also the reason, why I can't add series, movies or songs from unknown artists, of from sources I have. Because if the metadata server doesn't have those artists, the system can't add it... That honestly sucks. What are alternatives?
已收录讨论
Expand the replies to this comment to learn how AI was used in this post/project.
Classic Don't read and understand the post, but make a comment about someone I don't know anything about and hope that many will upvote it, so I can finally show my friends and family I archived something in life.
Ooo you got me!! Repeated my wording. The death of me! I’m so hurt, boo fucking hoo. I read your comments on here, arguing about how YOU thought the applications should work. Why not fucking read the application information before you install it. You’ll understand that it’s not for you. Then we won’t have to read your complaints about how it doesn’t fit your needs. And how the application should do x,y,z. To close, I ask: please link us to the GitHub request where you’ve asked the devs about your issue? or even better, since you’re such a smart hacker and diagnosed this “cloudfare” issue, where’s your PR for a change that can use the local data as you want? Yea I’ll be waiting.
Yea OP’s whole motivation on here is trying to create a problem that doesn’t exist. Because they think they are pointing out an issue. But it’s not an issue, it’s by design. **Just not a design that OP thinks is correct. TLDR: OP doesn’t like the app, but doesn’t want to build their own. Just here to complain.
Only Lidarr would ever need to query MusicBrainz. But no, Lidarr doesn't; that's what the upstream server OP is complaining about is avoiding.
Can it be switched to a locally hosted metadata server? I'm not super familiar with musicbrainz applications except for Picard, which I've used a handful of times.
They would be updating much more frequently. The only thing that matters is they are only ONE "user". They can update every 1 -5 minutes without any issue.
Because believe it or not, services like MusicBrainz are not super happy with the crazy amount of load that Lidarr usage without a centralised server would put on their infrastructure. If you use Headphones directly against musicbrainz.org, expect your IP address to get banned from all the MetaBrainz services pretty quickly. Bandwidth and compute have a cost.
What tinfoil hat? My request is still routed through cloudfare. A company I have nothing to do with. And if you read your link, they are collecting ASN, Country and Metro Area. That is called tracking. But be a smartpants here, please.
wait so arr stack tools all hit musicbrainz separately or what
I feel it very understandable and considerate that Sonarr, Radarr and Lidarr host a mirror of the metadata to avoid spamming the metadata providers. Metadata providers are an important part of selfhosted media. What I don’t get, is that they have no configurable metadata source in their administration settings. You can host your own MusicBrainz server for example. I should be able to hook my Lidarr up to that if I don’t want reliance on them. We saw that dependency very clearly went the Lidarr metadata mirror went down.
And this is exactly why I stopped using the full *arr stack for anything music related. The whole "you need our metadata server to even add a thing" model completely defeats the point of selfhosting. Lidarr's Spotify integration has been busted for ages and the 502s are just the tip of the iceberg when their API shits itself. For music, I switched to beets for metadata and just manage my library manually with Navidrome as the frontend. It's more work up front but at least I'm not praying to a Cloudflare endpoint every time I want to add an album.
The purpose of these tools is to manage your media. If you're importing your own content, they need to know how to identify it, so that they know what it is. If you want to manage your own media without these tools fetching metadata from somewhere, then, just... Put them in media folders on your NAS, or whatever you use.
“So every stupid thing I do is tracked somewhere else, out of my control, no option to opt out.” https://developers.cloudflare.com/network-error-logging/ Not sure if you need the tinfoil hat tbh
Classic “I CAN AND WANT TO DO IT ALL MY WAY” …. “But why doesn’t this app that’s widely used, that I didn’t create nor contribute to…. work the way I WANT IT TO?!?!” If you don’t like the way something is, help contribute to make it better, or create your own app.
Why is simple, they get the metadata from databases like tmdb or tvdb, but since they have so many users it would be very expensive to send all user requests to these external databases. So they download the entire database once and then update it like once a week or once a day. And all users requests go to their cache.
This is the tradeoff of a central metadata service . It smooths over a pile of upstream APIs that all behave a little differently, so the app team only has to maintain one translation layer instead of baking that logic into every client. The downside is exactly what you hit: it adds a dependency you don't control, and when that service is down, incomplete, or routed through Cloudflare, self-hosting feels a lot less self-contained. I'd want clear docs on what gets sent in those lookups and whether there's any way to bypass that layer, because the maintenance argument and the privacy concern are two separate problems.
I've yet to get around to looking at this: https://www.music-assistant.io/ But might be my plexamp/lidarr replacement
I don’t use the arr stack but got me thinking about the service I do use and whether the metadata service if being cynical aka telemetry collector.
You don’t necessarily **have** to use their metadata server. It’s possible to host your own metadata server. The setup can a little tricky but once it’s running it just works
without spof spotify rate limits ip instantly. proxy exists to protect developer keys.
It's the only way to avoid DDoSing small projects like Musicbrainz. The alternative is to host a metadata server yourself: https://github.com/metabrainz/musicbrainz-docker#recommended-hardwarevm Totally possible, but because it's a 100GB database, actively updated and therefore actively re-indexed, these are the system requirements: CPU: 16 threads (or 2 without indexed search), x86-64 architecture RAM: 16 GB (or 4 without indexed search) Disk Space: 350 GB (or 100 without indexed search) So, knock yourself out. (I say that sarcastically but I am actually dumb enough to be trying this myself).
I haven't been a fan of it for a long time. But there's very little you could do about it. They provide an insanely valuable service for free. There's no where else you can freely obtain the same information, easily. They are lucky people are willing to donate and we are lucky they are competent and run an excellent proxy/cache service. The only way it could ever be replaced, would be with an extremely similar service that would have the exact same "issues".
This is a very new app, but so far it’s working well. It visually looks very polished. A vast improvement from what your used to if you ask me. Has slskd and usenet integration. No need to rely upon Lidarr any longer. https://github.com/HabiRabbu/DroppedNeedle .
Just put it behind a vpn with gluetun it’s what I did