The push notifications are a centralized service that communicates with the Apple and Google push notification service. All those Element/XMPP/Zulip/etc apps run their own infrastructure that receives the notifications from your self-hosted server and passes it along to Google/Apple.
If you run an XMPP server and an iOS client for example you'll see in your logs that your server is connecting to, e.g. push.tigase.im or whatever if you're using Siskin. Another server for Monal. Another one for ChatSecure. and so on.
And they break. Monal for example rate-limits connections which leads to no push notifications for a while.
It's not unthinkable that they will either fold and follow the current trend or stop offering the service altogether, if it really does cost.
That's just plainly wrong.
A push notification will wake up the app for 30 seconds and in this timeframe enables it to receive all new messages directly from the xmpp server. When Monal receives an additional push message within these 30 seconds, the background time is increased by apple for another 30 seconds, resulting in a higher power consumption without any benefit.
That's why our pushservers ratelimits (lets better call it: deduplicates) pushes to 2 per 30 seconds.
Ceasing to operate our pushservers would be the same as ceasing to develop the app itself. And that would be bad even if no pushservers were needed.
So no, it is indeed unthinkable to stop offering these pushservers.
If you want development and the operation of the push servers to continue, you can always donate some money. Currently our funding especially for the push servers could be better…
If you run an XMPP server and an iOS client for example you'll see in your logs that your server is connecting to, e.g. push.tigase.im or whatever if you're using Siskin. Another server for Monal. Another one for ChatSecure. and so on.
And they break. Monal for example rate-limits connections which leads to no push notifications for a while.
It's not unthinkable that they will either fold and follow the current trend or stop offering the service altogether, if it really does cost.