[pilif]

If I was relying on a third party for push notifications (something I can easily do on my own, btw), the one single reason for downtime that would totally not be acceptable for me is them forgetting to renew their domain.

This is a huge indicator, IMHO, that they don't really believe in their service and don't even do the absolute minimum that's required for keeping it up.

Worse, depending on the type of message I'm passing through them, this might have privacy or security implications as messages will now be sent to a non-related third party.

Stuff like this is why I personally am very, very careful before outsourcing any part of my core infrastructure.

[pfg]

Yeah, because that could totally not happen to any serious organization[0][1][2].

Of course it's a major oversight, but these things happen. They'll learn from their mistake. I think it's wrong to dismiss a product just because they forgot to renew their domain, unless they keep messing up stuff like that all the time.

(I am in no way affiliated with Pusher)

[0]: http://www.theregister.co.uk/2003/11/06/microsoft_forgets_to...

[1]: http://www.webip.com.au/major-bank-forgets-to-renew-domain-n...

[2]: http://techcrunch.com/2010/03/27/foursquare-offline/

[EthanEtienne]

I here you, but sh!t happens. The lesson for me though is to choose your domain registrar as carefully as your infrastructure host, and have it be actively monitored like the rest of your infrastructure. I would still choose Pusher though. It's hard to scale everything in-house on limited resources, plus they make using WebSockets super easy (disclaimer: previous Pusher client).

[dxm]

We looked at using Pusher but having looked at their financials[0], we decided to implement similar functionality ourselves. Personally, I don't think the business is sustainable since it's trivial to build web socket functionality into an application, and hosting it on AWS or similar PaaS is cheap too.

[0] https://www.duedil.com/company/07489873/pusher-limited/finan...

[maxthelion]

Relying on this kind of financial data for an early stage company is always a bit risky. AFAICS, this is for the year ending March 2012, when we were significantly smaller than we currently are. We only turned on the ability to pay in August 2011!

Those financials have very little bearing on our current position. For the record, we're now profitable and growing healthily.