XMPP Service Operators - 2026-06-12


  1. Lilith [queer-spark.org]

    The only issue I really run into with having 7 day retention is with media in movim posts. I use nextcloud to work around that, but it's clunky and previews don't work correctly

  2. jjj333_p [pain.agency]

    https://downloadable.pain.agency/file_share/019eb9f9-84d6-715d-aaed-aa9c83dc46e5/27cfb8c1-1f15-4029-9678-81ad8bf4ee9b.png

  3. jjj333_p [pain.agency]

    i feel like im missing a message

  4. Lilith [queer-spark.org]

    https://queer-spark.org/file_share/019eba08-adba-76b6-9958-e6c507922270/c4eef074-2833-4561-8cc7-bbf52b30d180.png

  5. Lilith [queer-spark.org] retracted a previous message, but it's unsupported by your client.

  6. Lilith [queer-spark.org]

    Nah, you aren't missing anything

  7. Lilith [queer-spark.org]

    I was responding to what the TheCoffeMaker and jonas’ posted. What jonas’ was responding to I have no idea

  8. jjj333_p [pain.agency]

    ah alright

  9. jjj333_p [pain.agency]

    i have like 1 week mam but 4 week media 🤷‍♂️

  10. icebound.dev

    <iq type="get" id="834-10316" from="icebound.dev" to="xmpp.earth"><ping xmlns="urn:xmpp:ping"/></iq> <iq type="error" id="834-10316" to="icebound.dev" from="xmpp.earth"><ping xmlns="urn:xmpp:ping"/><error code="404" type="cancel"><remote-server-not-found xmlns="urn:ietf:params:xml:ns:xmpp-stanzas"/></error></iq>

  11. icebound.dev

    seems like XMPP.earth is down

  12. icebound.dev

    Sending server to server ping request to xmpp.earth Connection to xmpp.earth is closed; will attempt re-delivery once teardown completes. Inbound Server xmpp.earth authenticated using SASL mechanism EXTERNAL Successful server to server response received. Session is AUTHENTICATED. Successfully negotiated TLS connection.

  13. icebound.dev

    working again!

  14. icebound.dev

    not like it matters because *nobody can get a hold of xmpp.earth*

  15. Ilias (OpenIM.nl)

    What is an acceptable downtime, and how are you going to measure it. We cant expect 99% uptime from operators

  16. stratself

    > not like it matters because *nobody can get a hold of xmpp.earth* do they have a problem?

  17. stratself

    > What is an acceptable downtime, and how are you going to measure it. We cant expect 99% uptime from operators whatever makes their users happy /j

  18. Ilias (OpenIM.nl)

    I don't even promise 99% to my clients lol 😆

  19. icebound.dev

    > do they have a problem? stratself, I don't know, nobody can contact them

  20. icebound.dev

    > What is an acceptable downtime, and how are you going to measure it. We cant expect 99% uptime from operators Ilias (OpenIM.nl), there is no measurements of that, however this is exactly the issue with XMPP, there is very little professionally hosted providers, and there is absolutely no attempt to make any of it break even, its just always a massive expense someone within the community has to foot

  21. icebound.dev

    Conversations thankfully (to the best of my knowledge) break even or even profit from the sales of conversations on the play store, allowing them to fund their XMPP server

  22. icebound.dev

    > I don't even promise 99% to my clients lol 😆 so how do you convince people on the likes of Discord to hop away from Discord when nobody within the community can assure a decent quality of service?

  23. icebound.dev

    and you wonder why Matrix is being adopted much more than XMPP when it comes to IM for open source communities

  24. Menel

    Why do you think nobody can assure a decent quality? You say free provider *X* is bad, so *nobody* is good?

  25. Menel

    I see basically no downtime on conversations.im, yax.im, jabberfr.org. Conversations.im once had a uptime graphic for all connected servers. There were many with very good uptime

  26. Ilias (OpenIM.nl)

    I manage my free to use servers professionally. Uptime is currently 99%+. But I have my limits. I cannot guarantee 99% uptime because that means implementing HA, on call and monitoring+alerting, which all I cannot afford. I'm happy to do it all but who is going to cover the bill?

  27. Ilias (OpenIM.nl)

    > I see basically no downtime on conversations.im, yax.im, jabberfr.org. > Conversations.im once had a uptime graphic for all connected servers. There were many with very good uptime Also this. There are many free servers with great uptime

  28. stratself

    i dont think paid providers necessarily mean the services are better. maybe you bind them to a contractual obligation, but that's about it

  29. stratself

    also > it's just a massive expense i presume running a matrix server would be much more expensive than an xmpp one. yet people still do it, for free

  30. Menel

    The biggest expense is the time it costs

  31. Menel

    Especially the last two percent uptime

  32. Ilias (OpenIM.nl)

    Yup

  33. Ilias (OpenIM.nl)

    You can't change a culture in which everyone expects everything for free on the internet. If I decide to ask for a signup fee for the servers, i would have 0 users left. So i don't do it, so i don't promise 99% uptime.

  34. stratself

    > Why do you think nobody can assure a decent quality? > You say free provider *X* is bad, so *nobody* is good? maybe the problem is that free providers with good-quality are not very visible. though this is a problem for the curators of providers.xmpp.net, or any similar discoverability services

    🤔 1
  35. Menel

    Likely. That list doesn't have that in its metrics.

  36. stratself

    As a courtesy for potential users, I'd try to announce planned maintenance, report on unexpected downtimes, etc. Also make sure I'm reachable via xmpp and perhaps other avenues if problems arises. Maybe even have a MUC explicitly to discuss about the service itself but obviously those things are not reflected with the current metrics, yea

    ❤️ 1
  37. stratself

    (imo, admin reachability is a huge thing besides the technical stuff like uptime and stability. you can't always control if your server fucks up, but keeping members updated is a valuable thing)

  38. icebound.dev

    > Why do you think nobody can assure a decent quality? > You say free provider *X* is bad, so *nobody* is good? Menel, because i don't believe relying on non-profits provide a good enough service.

  39. icebound.dev

    and providers don't allow paid providers to be A tier, which I have rigidly opposed

  40. icebound.dev

    > Menel, because i don't believe relying on non-profits provide a good enough service. I am aware many people can provide an okay service, but you still have jobs, and you have no financial basis to keep it going, it comes out of your pocket and for some providers, if their donations drop, their service disappears.

  41. icebound.dev

    > I see basically no downtime on conversations.im, yax.im, jabberfr.org. > Conversations.im once had a uptime graphic for all connected servers. There were many with very good uptime Menel, afaik conversations makes a profit from the play store, after expenses

  42. icebound.dev

    > i dont think paid providers necessarily mean the services are better. maybe you bind them to a contractual obligation, but that's about it and when you rely on a provider for your daily life, would you not like a contractual obligation to ensure that you get a minimum service, or get financial compensation for it?

  43. icebound.dev

    > I manage my free to use servers professionally. Uptime is currently 99%+. But I have my limits. I cannot guarantee 99% uptime because that means implementing HA, on call and monitoring+alerting, which all I cannot afford. I'm happy to do it all but who is going to cover the bill? exactly.

  44. icebound.dev

    > The biggest expense is the time it costs exactly, which is why profit (dare I say the evil p word) compensates you for your time

  45. jjj333_p [pain.agency]

    > so how do you convince people on the likes of Discord to hop away from Discord when nobody within the community can assure a decent quality of service? you say that like its not a literal meme on discord how often the service goes down (or at least it was like 5 years ago when i still used it much)

  46. icebound.dev

    > You can't change a culture in which everyone expects everything for free on the internet. > > If I decide to ask for a signup fee for the servers, i would have 0 users left. So i don't do it, so i don't promise 99% uptime. I do believe there is a good enough market for this, snikket and conversations both do paid hosting

  47. icebound.dev

    > maybe the problem is that free providers with good-quality are not very visible. though this is a problem for the curators of providers.xmpp.net, or any similar discoverability services I do believe this is true too, I feel like there is a few bad apples in the providers, which are A or B ranked, and then this is what is causing so many issues. A provider like conversations.im is listed as D tier because it is not providing the information that providers wants, rather than them being D tier for a poor service.

  48. icebound.dev

    providers is solely about compliance, not about the quality of service.

  49. icebound.dev

    > you say that like its not a literal meme on discord how often the service goes down (or at least it was like 5 years ago when i still used it much) jjj333_p [pain.agency], and yet I still have had so many people tell me that XMPP is less reliable because x and y provider keeps going down

  50. icebound.dev

    providers ranked A or B

  51. icebound.dev

    07f.de is the single biggest provider I have had issues reported to me

  52. Ilias (OpenIM.nl)

    >> You can't change a culture in which everyone expects everything for free on the internet. >> >> If I decide to ask for a signup fee for the servers, i would have 0 users left. So i don't do it, so i don't promise 99% uptime. > I do believe there is a good enough market for this, snikket and conversations both do paid hosting Maybe worth exploring that market

  53. jjj333_p [pain.agency]

    idk, but you throwing a fit isnt gonna help a lot

  54. icebound.dev

    and after months, i still haven't heard anything back after trying to reach out

  55. icebound.dev

    > idk, but you throwing a fit isnt gonna help a lot I am doing what I can lol

  56. icebound.dev

    and I am not throwing a fit

  57. jjj333_p [pain.agency]

    throwing a fit is all you can?

  58. jjj333_p [pain.agency]

    > and I am not throwing a fit yes you are

  59. icebound.dev

    no I am not

  60. icebound.dev

    I have called a operator meetup exactly due to these concerns. I am working on a UK provider, exactly due to my concerns here.

  61. icebound.dev

    I have worked hard on outreach and it feels like right now the operators are the bottleneck, I can get people interested, but the moment it comes to finding a host for their account, thats where I lose people

  62. MattJ

    There are projects that work on that problem (providers.xmpp.net, etc.)

  63. icebound.dev

    in fact, I have seen a few of those go for paid hosting from conversations.im because of their bad experiences with public providers.

  64. icebound.dev

    > There are projects that work on that problem (providers.xmpp.net, etc.) providers is part of the problem :/

  65. MattJ

    If they are insufficient somehow, contribute or fork?

  66. icebound.dev

    and quite frankly I am giving up trying to work with them...

  67. icebound.dev

    they are rigid on the idea of discouraging paid providers, and there is other controversies, such as the one holger brought up

  68. icebound.dev

    > and quite frankly I am giving up trying to work with them... on the verge*

  69. stratself

    >> Menel, because i don't believe relying on non-profits provide a good enough service. > I am aware many people can provide an okay service, but you still have jobs, and you have no financial basis to keep it going, it comes out of your pocket and for some providers, if their donations drop, their service disappears. again i see this as a problem with the providers' list not mentioning the expense or sustainability or whatever other factors imo if you recommend xmpp to your peers, you recommend the servers too. dont give them choice paralysis

  70. MattJ

    I guess my underlying point is: you can't ever force every XMPP service to always have 100% (or even 99% or even 90%) uptime

  71. icebound.dev

    I feel like what providers does, and what the operators do are two separate things, and the providers demand that you use *their* muc, yet all the operators which are having issues are here...

  72. MattJ

    So the problem shifts towards recommending/promoting servers that are good

  73. icebound.dev

    > I guess my underlying point is: you can't ever force every XMPP service to always have 100% (or even 99% or even 90%) uptime of course not, but people funnel through providers, sign up to the likes of xmmp.earth and 07f.de which I and many others still can't get into contact with the admins of

  74. icebound.dev

    and then they are discouraged due to downtime, or broken features... and nobody here can help out, because what we meant to do?

  75. stratself

    tbh just make a new providers list at this point

  76. icebound.dev

    > So the problem shifts towards recommending/promoting servers that are good but the people who are doing this, are difficult to say the least to even discuss these problems with

  77. icebound.dev

    when I try to discuss it I get hit with "we worked really hard on this" and it makes me feel bad

  78. MattJ

    It's all open source and open data... we have tried and tested solutions to this problem :)

  79. MattJ

    If you find problems with their approach and they are happy with how things are, you can make an alternative

  80. icebound.dev

    I have an idea, but it will take time to implement, and I am already swamped :)

  81. MattJ

    Right

  82. icebound.dev

    > If you find problems with their approach and they are happy with how things are, you can make an alternative I have wanted to use the providers API and do some monitoring on some of the providers to see the scale of the problem for a while, just I cant find the time for it

  83. MattJ

    And I'm not saying it's not a good thing to try and help and encourage operators to improve their reliability (e.g. by sharing best practices, or whatever they need)

  84. MattJ

    But it is not a problem that will ever be solved

  85. MattJ

    It's similar to wanting every client to implement the same XEPs up to the same standard. Nothing you do in an open ecosystem can stop someone writing a new half-baked client over the weekend, publishing it, and then effectively abandoning it.

  86. stratself

    >> i dont think paid providers necessarily mean the services are better. maybe you bind them to a contractual obligation, but that's about it > and when you rely on a provider for your daily life, would you not like a contractual obligation to ensure that you get a minimum service, or get financial compensation for it? also, i'd rather pool in resources with my family/friends/communities and set up our own server(s). if we go down we go down together like a team like i respect the option but it's not always the most appealing yknow

  87. MattJ

    It's just a downside of open ecosystems that we have to recognise, and work around

  88. icebound.dev

    the thing is, if I encourage people to use a specific server when I outreach, we become matrix, if I funnel them towards providers, many get turned off when they hit a bad operator ranked highly on the list. Without dedicating even more time to then do my own monitoring on providers and fish out the bad ones, it feels like attempts at outreach are just being wasted.

  89. stratself

    do you hate matrix or do you want stability?

  90. icebound.dev

    > also, i'd rather pool in resources with my family/friends/communities and set up our own server(s). if we go down we go down together like a team > > like i respect the option but it's not always the most appealing yknow not everyone has the ability do it with good uptime and good security, that is why services exist.

  91. icebound.dev

    > It's similar to wanting every client to implement the same XEPs up to the same standard. Nothing you do in an open ecosystem can stop someone writing a new half-baked client over the weekend, publishing it, and then effectively abandoning it. The is true, but the other thing which has been discussed before is the fact the providers use the xmpp.net domain, this is not the xmpp.org domain but its so close that it looks like an official project.

  92. MattJ

    You're welcome to a subdomain of xmpp.net when your project is ready :)

  93. icebound.dev

    > You're welcome to a subdomain of xmpp.net when your project is ready :) thats not my point

  94. icebound.dev

    My point is, it looks like the official thing, and thus when people have issues it reflects badly. Meanwhile if someone uses say conversations and they had a bad experience they blame the app.

  95. icebound.dev

    its not quite the same...

  96. MattJ

    Again, that can't really be helped. People register xmpp.<TLD> domains that aren't official, but I'm sure some people think they are somehow "official" (whatever that means)

  97. icebound.dev

    I still believe that the lack of paid providers is another issue...

  98. icebound.dev

    > Again, that can't really be helped. People register xmpp.<TLD> domains that aren't official, but I'm sure some people think they are somehow "official" (whatever that means) this is why companies buy out all the major tlds for their company :p

  99. icebound.dev

    but yeah thats expensive.

  100. icebound.dev

    > I still believe that the lack of paid providers is another issue... people here and in the providers seem to disagree quite firmly on this :)

  101. MattJ

    As a paid provider, I can't comment :P

  102. icebound.dev

    People complain about not having the time to dedicate to provides, which the solution to this is you getting paid for your time, but then thats also a bad option for some reason?!!?

  103. icebound.dev

    I just dont get it

  104. icebound.dev

    > As a paid provider, I can't comment :P yeah you are one of the few :p

  105. stratself

    i dont wanna say paid providers are inferior. but perhaps they could use an alternative channel of advertisement

  106. stratself

    if you want a mixed channel where both are accepted, you can do it on your own

  107. Menel

    We have multibe for profit providers and have many many free ones. I don't see the issue that's all

  108. icebound.dev

    > We have multibe for profit providers and have many many free ones. > I don't see the issue that's all I know of two paid providers which are well known.

  109. icebound.dev

    out of a list of what, 80? free providers

  110. Menel

    And of course you should recommend exactly one server to your friends. The same you use in the best case. That doesn't make it centralised. Like I invite them all to my own server. Making the ecosystem less centralised in fact

  111. icebound.dev

    > And of course you should recommend exactly one server to your friends. The same you use in the best case. > That doesn't make it centralised. > Like I invite them all to my own server. Making the ecosystem less centralised in fact im not talking about my friends, I already onboarded them, when you do talks on XMPP, when you go to events and are trying to convince people to join the network, you dont want them all going to one server

  112. Menel

    > out of a list of what, 80? free providers Yes. What do you expect if one can setup an instance on a vps in one afternoon

  113. icebound.dev

    > Yes. What do you expect if one can setup an instance on a vps in one afternoon yes but then they follow compliance, and become B tier, they are then listed in clients for people to use, people sign up to them and then 1 year later when the free provider burns out or runs out of funds, disappears.

  114. stratself

    >> And of course you should recommend exactly one server to your friends. The same you use in the best case. >> That doesn't make it centralised. >> Like I invite them all to my own server. Making the ecosystem less centralised in fact > im not talking about my friends, I already onboarded them, when you do talks on XMPP, when you go to events and are trying to convince people to join the network, you dont want them all going to one server list 5 that you like. be opinionated

  115. Menel

    icebound.dev: then give a list of three, and mention where they can singup for a fee.

  116. icebound.dev

    > icebound.dev: then give a list of three, and mention where they can singup for a fee. at this point I have honestly just recommended people to pay conversations.im/snikket

  117. Menel

    You can continue to rant about for free services someone lists somewhere. But best do it in their room, where they are listed. It's nothing we can do about providers here

  118. icebound.dev

    I was rather blunt at the last talk I did at a small meetup, and pointed out that providers are hit or miss, if you really want to rely on it, pay for it

  119. Menel

    Personally ai don't recommend the providers list anywhere

  120. Menel

    Personally I don't recommend the providers list anywhere

  121. icebound.dev

    > You can continue to rant about for free services someone lists somewhere. But best do it in their room, where they are listed. It's nothing we can do about providers here I kinda derailed my original point of speaking here today, which was xmpp.earth is down and nobody can reach the provider still...

  122. icebound.dev

    and people sign up to xmpp.earth from providers

  123. stratself

    now

  124. stratself

    who led people to providers.xmpp.net?

  125. icebound.dev

    > who led people to providers.xmpp.net? we hand out provider leaflets at conferences

  126. stratself

    why are they being led there? did someone hold them at gunpoint and tell them to signup?

  127. stratself

    >> who led people to providers.xmpp.net? > we hand out provider leaflets at conferences oh how about you just don't do that

  128. Menel

    This isn't the providers channel.

  129. icebound.dev

    no its the main point at onboarding, and many clients now are integrated into providers, take gajim, it gets its server list from providers, filters the A and B out and sorts by name

  130. icebound.dev

    > This isn't the providers channel. well this is what I said above, there is no communication between providers and operators, when they go hand in hand

  131. icebound.dev

    again this is exactly why I am working on an operator meetup online, because it all feels fragmented

  132. Menel

    You yourself estsblied they do what they want and you don't like their ranking. So maybe stop using them

  133. Menel

    It's nothing anyone here can do beside stop using it

  134. icebound.dev

    > You yourself estsblied they do what they want and you don't like their ranking. So maybe stop using them I remember what it was like before providers, just because providers is causing issues doesn't mean I want to abandon them entirely and start over.

  135. icebound.dev

    instead I feel its important providers and operators are better connected

  136. Menel

    I still don't we what you think will change. You yourself said you tried again and again. I would prefer if you would post it all in the providers channel not here. Because only they can do something about their list. No operator can.

  137. Menel

    I still don't see what you think will change. You yourself said you tried again and again. I would prefer if you would post it all in the providers channel not here. Because only they can do something about their list. No operator can.

  138. icebound.dev

    Honestly I dont know

  139. icebound.dev

    its just that feeling of frustration, that you throw your time at something and it feels little changes