Meteor.com free hosting ends March 25, 2016

I respect the decision to charge money for a service, but why would you give just two weeks notice when your public docs only recently said this would be free forever? And no tools to help us estimate the cost/month should we move existing simple apps to one of the Galaxy services (e.g., I have no idea what $.0035/hour means for my small volume app - does that mean < $2.50/month since there is some way to know < 1 instance at a time runs?)

Two weeks notice (less for me - I just checked my email today to find I have only 11 days - and I feel bad for the poor bastards who are still on vacation and don’t yet know) is not a reasonable amount of time, will be very disruptive, and you are going to loose a chunk of your community if you follow through with this. I’m really hoping you change your timelines and add some planning tools to help existing customers make an informed decision.

On a personal note, I just built my first ever web site on Meteor (for about 40 of us that play in a racquetball ladder) and it’s been working fine for the last few months (save occasional short-term Mongo outages on meteor.com), but this is a side-project for me and not sure if/when I’ll have time to try to migrate it in the next 10 days. And I know there are larger scale players that have bigger bets on the free tier (e.g., I’m just finishing up a series of Coursera classes on Meteor development which use the free site quite heavily for student created peer-graded assignments).

So what prompted this abrupt timeline for change? Is MDG in financial trouble?

4 Likes

The far more disturbing thing about this whole discussion is actually not even anymore the announcement itself. I find it more disturbing that within 3 days of the announcement nobody from MDG has thought it would be a good idea to respond and address all concerns properly.

@sashko talked about this in one of the transmissions that he dislikes these articles with shitty headlines that put meteor in a bad light but then you’re doing it yourself. The damage has been done but instead of fixing it you don’t clarify and respond to the community within 3 days. You can’t be surprised people write shit about MDG on hackernews if you can’t reply about one of the most major core decisions of your business within 3 days.

9 Likes

I don’t have much to add to this conversation as I’ve only hosted using *.meteor.com once.

I’ve been experimenting with creating packages and stuff, and as everyone knows, going to atmosphere and checking out a package at packageName.meteor.com is for sure the most satisfying and helpful tool for developers to check out the usage of the package.

Besides that though, I don’t mind the removal of the free hosting. I think I took it for granted when it obviously hasn’t been good business practice for MDG.

If they can find a way to allow package developers to still free host, I think a large chunk of Meteor users will still be happy

2 Likes

Just posting to formally register my agreement that giving people two weeks notice is pretty crappy given previous proclamations that free hosting will always exist. I understand that businesses need to change to survive but are MDG haemorrhaging money that fast that two weeks notice is all they can give before flicking the off switch?

If MDG make this type of massive change at very short notice for free hosting it makes me worry about hosting my apps on Galaxy for fear of getting a polite letter after 6 months telling me my bill will be tripling in two weeks time, signed off with “Thanks for using Meteor and best of luck on migrating your projects from Galaxy”!

1 Like

I personally would like to know the reasoning for the short notice

3 Likes

I agree, and I won’t bother with my theories because that’s all they are, theories. I will say that I support the general decision, but not the short notice. I would also say that hopefully they’ll put the data on ice for a while while they sort out what decisions are being made regarding Community package docs, demos and other important websites on *.meteor.com

3 Likes

For what it’s worth, I moved one of my apps from free sites to my Webfaction server and the data to mLab this weekend. It was fairly simple once I found the right mongoexport command to use. (Granted, that took a while, but mostly because I’m new to much of this.)

It was essentially this:

  • installing Mongo on my local machine
  • meteor mongo --url my-app.meteor.com to get the username, password, url, port, and database name (username:password@url:port/databasename)
  • save the data onto my local computer using mongodump --host hostname:port --db databasename -u username -p password -o ~/Desktop
    *package up my app into a tarball using meteor build --architecture os.linux.x86_64 ./in my local machine’s directory
    *upload the .tar to my server through ftp and unpack with tar -xzf my-app.tar.gz
    *set local variables, including the MONGO url using the mLab uri
    *entering the /bundle/programs/server/packages directory and running npm install
    *going back to the /bundle directory and running nohup node main.js (I know this would be better served by a cron job, but I’m not confident enough in my knowledge to share how to do that here)

There are a couple Webfaction specific steps beyond that, but they aren’t too much hassle.

I just wanted to share my process in case it helps someone else - I’m sure there are more elegant ways to do some of this through a script, but that’s not in my knowledge base at this point…

2 Likes

Hi,

I’m very disappointed with the choices and way it was communicated. There was a previous commitment made that I banked on when learning meteor that the simple deployment option would never go away. How can MDG expect people to put trust in them when they are throwing their own words out the window?

2 weeks to migrate 10,000s of apps? No tier at less than $25/month? I can get an good sized VPS for less than $10/month, why does app hosting have to be more than twice the cost? The entire thing reeks of desperation to me - is MDG that close to the wire money wise that they were forced into this position? Forced into trying to convert some of those free users into paying customers?

I’m not sure, I just know that have a simple way to deploy apps is important, a commitment was clearly stated and has been walked away from. The final solution to this is extremely abrupt and seems force everyone down a path that no one would ever even expected to have to consider.

Mark

3 Likes

I’d guess that the actual number of apps in need of migration is much, much smaller. Most of these 10,000s are Hello World projects or dropped stuff.

4 Likes

I second this opinion.
Although having just started learning meteor, I put a lot of time and effort in evaluating different frameworks for my startup before deciding on meteor.
Changing a fundamental pillar of a framework/community on such short notice and in such degree (from free forever to one month worth of credit) is alarming and makes me second guess my decision.
I can understand that there might be legit business driven reasons to change services.
And yes you should not rely on free services to fund your own business.
However for me to rely on a technology I need a reliable technology provider. Behaving in such a way is not what I would call reliable.
But I have to cut MDG some slack. Meteor and MDG are very young compared to other frameworks or providers. They created a great thing and got a lot done in an incredible short time span. And I love it. I love to use it and I love to watch it grow and change. Probably RoR or Heroku never had such problems but they never had this kind of traction at such an early stage (Disclaimer: to “young” to really know).
However creating such a great product gets you a lot of attention and publicity. This affects the way you are able to change the way you do things.
You start of with great ideas and then learn that there are better solutions to what you came up with in the first place. So as every good student would do you move to the next level. Breaking changes will always hurt but not doing so will most likely hurt you even more.
You should not try to preserve your original state. You need to move forward!!
But please be aware that with this kind of traction so early in the game it affects a lot of people. So you need to adopt to this situation.
This gets me to my point (sorry for the long rant):

  • I would totally be okay with such changes if I can understand the reasoning behind it (which I can’t atm)
  • I am able to adopt to the new situation in a reasonable time.

P.S. reasonable time for a free service might be debatable, but for a such a pillar of the community 2 weeks are really really short. Just one example coursera.org started a specialization 2 months ago which includes deployed meteor apps as assignments. At least one of the classes runs past the 25th. Changing these classes right in the middle of it might prove very difficult.

2 Likes

My .02 is that, ever since the “Is Meteor Really Open Source?” thread, @sashko has really gone above and beyond to communicate clearly with the community about what the MDG group goals are.

What he’s doing is absolutely vital if you have any chance of convincing developers to adopt your platform, especially considering Meteor’s stack isn’t aging too terribly well and if I’m going to spend 6 months of my professional life working with clients on a specific platform, I need to know if I’m setting this client up for a world of hurt (read: don’t expect to grow or have access to updated versions of the server and database). Because of @Sashko now, and with Transmission, I can say this fear as been alleviated. MDG is growing their product by not ignoring the industry, but embracing it.

These changes imply that MDG really care about devs. Because seriously, aren’t we their market?

ALL THAT SAID however, it doesn’t seem like he influences policy per se — please note that this is just rank speculation on my part and maybe irresponsible of me to suggest. But the reason I say that is because two weeks notice before spinning down the free meteor sites is about as deaf and obtuse a decision as I can possibly imagine (well, sure companies have made more anti-dev decisions than this, but they’re usually never around long enough to fix them).

I really don’t care anything about the free Meteor sites that launched, had 5 hits and then lay dormant for 2 years. But this isn’t what we’re talking about. We’re talking about Atmosphere’s seeming reliance on the free Meteor tier to host proof-of-concepts and published packages. Packages that WE. ALL. USE. WE. ALL. RELY. ON.

Would it really have been that much more expensive and complicated to say, “if your free Meteor deployment hasn’t received any traffic in 6 months, we’re nuking it in 2 weeks. If your free Meteor deployment has received traffic these last couple months (ie most likely an atmosphere package), then you get 90 days.”

Because seriously, I’m not at all advocating that you even maintain the free tier. This decision is simply ANTI DEVELOPER, as far as I can tell — 14 days sunset on a product you promised never to eliminate? Seriously?

I love this stack and I love this community and I want nothing but the best for it, so I’d rather express my opinion about this than keep silent.

3 Likes

I’m the tech lead on Apollo, and help out with open source Meteor stuff as well - in those concerns I’m directly responsible.

5 Likes

The short notice implies that you’re not treating your users with respect. I didn’t expect that.
Why should I entrust anything to you if you could not handle something as simple as this more gracefully?

As a company, that is. I respect the open source work, and all the talent you have on your team.
But this decision is poorly executed.

2 Likes

Maxhodges has replied, with the stated offering as it was communicated previously. Why no comment? It makes MDG look foolish and is akin to burying ones head in the sand…

Given this is a forum, the lack of response around this important topic is frustrating and doesn’t reflect well on the company - I expected more.

The argument was that:

[quote=“prime, post:127, topic:19308, full:true”]
MDG didn’t lie. meteor deploy will never going away. After March 25, 2016 meteor deploy will serve Galaxy :grinning: [/quote]

However that’s not true (and I thought it was said ironically, hence the grinning smiley at the end). The blog post said “free hosting to every Meteor developer through our meteor deploy feature”, so they really meant “free hosting”. And then it also went on to say that they “already started work on that work so that every developer can use Galaxy free of charge for simple apps”.

I know the MDG team is probably tired of this discussion, but I think that to not lose people’s trust they should finally stand up and say something along lines of “yes, we screwed up, we said that we will ensure that free hosting will not go away, but unfortunately we couldn’t fulfill that promise. However this is what we will do instead to satisfy everyone: …”. This would finally end this discussion and would let us focus on how to actually move forward.

5 Likes

@marktrang: Will you provide ways to configure redirects from the *.meteor.com? One useful thing was that it looked much better for packages to be hosted under *.meteor.com.

4 Likes

I was thinking the same thing. Wasn’t even worried so much about how things looked; just about lots and lots of broken links.

2 Likes

Meteor Operation Room

Day 0:
rohit2b: Sir, operation Kill Free Bird is ready. Shall we execute?
gschmidt: Has any word of the operation been leaked to the community?
rohit2b: No sir, they will be totally taken by surprise.
gschmidt: Perfect! GO

Day 1:
gschmidt: rohit2b, status report on Kill Free Bird.
rohit2b: Sir, there are rumors we said the free tier was never going away.
gschmidt: Have sashko deny immediately.
rohit2b: Done sir!
… sometime later
rohit2b: apparently there is photographic evidence we said the free tier was never going away.
gschmidt: Can we photoshop this “evidence”?
rohit2b: No I don’t think so.
gschmidt: DAMN… well at least we gave them generous amount of time to migrate, that should ease the blowback
rohit2b: Right you are sir.

Day 2:
gschmidt: community report!
rohit2b: They are still complaining sir, and they are saying that 2 weeks is not a reasonable amount of time. How should we respond?
gschmidt: Ingrates!, no response for now, this will blow over quickly, it always does

Day 3:
rohit2b: Sir, continued complaints about the migration timeframe and broken promise, but now they are also complaining about us not responding
gschmidt: damn, I thought this would have blown over by now…
rohit2b: how do you wish to respond?
gschmidt: We cannot respond. If we do, then we’ll have to respond to every post.
rohit2b: but sir, this is easily the most viewed topic of the past week, surely we can just resp…
gschmidt: SILENCE

Day 4:
rohit2b: still complaining sir…
gschmidt: Do they never tire?
rohit2b: what if we just said we’ve heard the response but just financially can’t make any other choice?
gschmidt: no,no, they known we’re well funded… we need a different strategy
rohit2b: how about if we tell them that it would take weeks to formulate a response, and by that time it won’t matter anyway, so that’s why we can’t respond
gschmidt: hmmm, I like where your head is at, but I think I have a better strategy
rohit2b: perfect!, what is it
gschmidt: how about… we… give them no response! yes, let’s do that

:grin: all in good fun

19 Likes

We are exploring easy ways to keeping the *.meteor.com for specific package demo apps (see that thread) but cannot deliver this wholesale to all *.meteor.com apps for a variety of liability reasons.

4 Likes

Hey great news - we’re going to talk about this on the next TRANSMISSION show, with the Meteor and Galaxy project managers, @zoltan and @rohit2b. Show is recording on Friday and will come out next week! Get notified when the show comes out, and help ask questions here:

2 Likes