I think it's kind of weird to see Google/Firebase team members swarming around in dozens on HN when they present a new feature, which seems to me like a coordinated marketing effort by Google. But once problems arise or they get criticized by their own users, no one seems to be available for a comment.
I was thinking the same thing. There's a giant circle of patting each other on the back every time they roll something out, and always have that very precise disclaimer that they work for the big G.
Then again, as a meta comment I will point out that the fact that both you and I so clearly notice this suggest that we are following these updates much more closely then most of the folks here.
True, I used to be really interested in Firebase and their set of products. But because of several pretty bad experiences related to vendor lock-in and (very) high prices I switched over to an open source alternative.
It might not be so well known, but I found http://www.deepstream.io to have a relatively similar feature set, without having to deal with bad support (if it bothers you, just fix it yourself) and high prices, since you are the one who picks what suits you best.
I was about to post the same thing. Of the two Google cloud stories I saw on here yesterday, it felt like one in three posts had "disclaimer: I work on the Google cloud team" or something to that effect. As of this writing, there's 100 comments over 4 hours and not a peep from any of them that I see.
Sometimes I am wondering who is actually steering the discussion to a certain point. I wouldn't go so far to call it censorship, but if the Google employees take over the comment section, it is very hard to call the flow of information "unbiased".
YCombinator shouldn't let companies use Hacker News as a marketing tool, at least not to the extent Google does.
Or it's just engineers that swarm around since that's the demographic here but this needs a response from PR, not from engineering. What do you want google/firebase engineers to say?
The engineering response is pretty clear-cut. Polling every minute from an uncached SSL connection is expensive, don't do that. That's pretty obvious and the author admits this was a mistake. The complaint is about everything that's not the engineering part - support & communication.
I'm sorry, I can have 50'000 devices poll every minute via an uncached SSL connection on my 16$/month server from Online.net, including bandwidth and traffic costs.
I'm not sure what you're smoking, but I want that, too.
Seriously, at Google, TLS is terminated in hardware, and bandwidth costs less than a cent per terabyte for ISPs or Google.
The costs that this caused are basically nil.
(This, btw, is why I'm using kubernetes on rented bare metal, paying 30$ a month, for what would cost me, including traffic, around 900$ (Google Cloud with rented servers) to 63'000$ (Firebase) a month.)
