Remove dead instances from list

[Darin755]

Describe the feature

Currently there are quite a few dead instances listed under instances page. Here are a few examples:

https://watchapi.whatever.social

https://pipedapi.qdi.fi/

https://pipedapi.colinslegacy.com/

https://pipedapi.tokhmi.xyz/

All of these do not work and have not worked for a extended period of time. Would it be possible to go through and remove the ones that are dead? I would open a PR but that is not possible for a Wiki page.

Why would this be useful to add?

Right now if someone randomly clicks on a piped instance there is a good chance they will find one that does not work. We should clean up the list so that it is easy to find alternative instances now that Youtube is breaking them randomly.

Concept(s)

perhaps go though each one and visit the site. If you are met with infinite loading, not found or similar error it is likely due to the instance mo longer existing. Theoretically this could be automated with a bot that updates the Wiki.

Additional context

#615 might be applicable here as this is an issue moving the instance list to a public repo. It is also important to note that if someone were to give up there DNS domain a bad actor could buy it and do something nefarious. I think that is unlikely but not impossible.

Acknowledgements

• I have searched the existing issues and this is NOT a duplicate or related to another open issue.
• I have written a short but informative title.
• I filled out all of the requested information in this form.

[opusforlife2]

Syncpundit works. Instances go down and come back online all the time.

[ghost]

One of the instances from the list (I don't remember which) never worked for me, and then I found out it was blocking IPs from my country. It worked just fine through a proxy service.

So, before reporting an instance as dead, please test it with proxies or VPNs from different regions.

[Darin755]

I removed the only that came back up. However, I think instances that block users based on countries should be treated as down. If they want to do that that's fine but don't put it on the public list or at least label it as restricted.

With that being said I think that doesn't apply to most instances. Maybe a not could be setup to remove instances that are down for a longer period of time. It could wait a week or two for it to come back up and if it doesn't it could remove it from the list. If a instance admin wants it added again they can open a PR.

I think that for the best experience we need a list of instances. I am thinking about making my own with a bot that updates it.

[opusforlife2]

I suppose there could simply be a toggle that switches instances silently whenever the currently selected one goes down.

[BobbyMcJefferson]

As of today, videos no longer load and I can no longer select an instance because the list has disappeared.

edit: I guess it's back now, it wasn't there a few hours ago.

[Bnyro]

Right now if someone randomly clicks on a piped instance there is a good chance they will find one that does not work. We should clean up the list so that it is easy to find alternative instances now that Youtube is breaking them randomly.

We don't have any way to tell if the site is temporarily or permanently down / no longer providing Piped as a public service, so how do you want to know which ones to remove and which not? As long as the instance admin doesn't open an issue to remove the instance, we pretty much don't have any chance.

The instances list from the wiki is not the recommended way to choose instances, the list at https://piped.video/preferences only contains instances that are up and working, see https://github.com/TeamPiped/instances-api.

[opusforlife2]

It even has an Uptime column to help user know which instances are more reliably available at any given time.