Galaxy account suspended without notice

I’m sorry I have to revert to this but given the usual Galaxy support takes several hours I see no other option than to post here in addition to having contacted their support and billing support email addresses.

I woke up this morning to a message from my uptime monitoring service that the services I host on galaxy are down. I had an annual pre-paid plan last year, but due to client’s needs having some variation in peak loads decided to cancel this reverting back to pay as you go hosting.

I received confirmation that my annual subscription was cancelled on 6th of December, expiring on the 10th.
I received an invoice for additional servers I have been running with payment confirmation on the 5th of December.

Logging in to Galaxy just now I’m told my service is suspended an of 4:40am UK time on the 12th of December.

Trying to restart my containers I see this is due to an overdue balance. I had absolutely no communication from Galaxy to this effect. I’m running two business-critical applications on my account so this is absolutely time critical.

Any recommendation how to get fast support from Galaxy to sort this out is welcome. Both applications affect users in Europe, so I cannot really wait until people on the US East coast are starting to work.

Update

19 hours have passed and I’ve just been notified by the support that they ‘think’ my service is activated again.

  • No form of apology.
  • No clarification of the reason for the outage.
  • Only: Your issue should be resolved. Please let us know if you still have issue (sic). Thanks for using Galaxy

Thank you to all those who made recommendations for alternative hosting. I’ve started the process of moving my products off Galaxy and will likely use Heroku or AWS moving forward. This has been an eye opener as to how badly Galaxy is supported when they simply suspend your account with no reason or accountability.

Update 2

A day after the first update I had confirmation that this was human error by Galaxy and a brief apology. There was no advice how this could have been avoided through different comms channels or suggestion that they will review their process that led to this error. So the RFO I will sent to my clients will remain pretty bare.

I was also offered a refund of my previous month’s cost (given I had a 1GB annual subscription this was far less than the actual cost for that month.)

My request to be put onto a 1 hour SLA support (I really don’t use their support much so don’t expect this to generate cost or benefit me unless I experience a similar outage) remained unanswered.

My recommendation to clients will be mixed. I still like the ‘don’t worry about it’ approach whilst a VPS that I run elsewhere will likely take me more time to manage. But I will also put an emergency solution in place to:

  • Make sure I have always an up to date deployment process to a VPS using the live database if needed.
  • Make sure I can update DNS records faster (this is not Galaxy’s problem but caused me issues as I did not control the client’s DNS records) and they had a 24h TTL setting when I finally got through to their IT team.
2 Likes

This is quite unfortunate. I think the method you are using is atm the fastest. The forum is visited randomly by staff members so I would expect them to be here at the very least around 3pm ish :confused:

Is it possible to temporarily migrate stuff to a different account?

I’m just deploying with meteor up to a VPS that costs a fraction of the Galaxy price. I’m using those for staging and new releases so at least it’ll be fast to spin these up.

The trouble is the IT supplier managing the DNS for my clients in very very slow to respond and has a 24h TTL on the records, so my client will need to notify customers of a URL change.

I had 99.998% uptime with Galaxy, the VPS based uptime was at 99.95% but my reason so far has been that Galaxy is a managed system and I’m a one man band developer running stuff that needs to be live no matter if I’m in bed, on a flight, or just have an evening out.

Given they don’t have policies in place to at least confirm a planned action or notify a user when they suspend an account, I will need to see how quickly they can resolve this before I decide if I’m still planning to pay 8x as much as a comparable VPS costs me.

1 Like

Makes sense… Any response already?

Not yet (5h after raising this via email). I’d presume support would weigh in on here to show they are proactive as well.

So it’s nearly 12 hours since my account has been suspended.

  • I have not received any notification as to why it would be suspended.
  • I have not received any advance notification that this would happen or notifications of outstanding payments.
  • I have checked with another former client who still has an account attached to Galaxy, which I have access to that he is not in arrears either.
  • The account was suspended after business hours PST time. So even if there was not an issue with my time zone I would have been unable to do anything about it for over 12 hours.

Galaxy, I pay $118.80 per month for 3 containers that cost on alternative providers $10.50. I’m really struggling to understand how you can justify this level of service, lack of response given the cost.

Please get back to me here or on my support ticket. I’ve had a 12 hour outage now on a service that I have sold my clients on the basis that it has been very reliable. You have now dropped the monthly uptime to less than 98.4% And there still seems to be no way to get hold of anyone.

At this point even if I want to continue using you I have at least one client who won’t be willing to pay this bill and is requesting a refund.

1 Like

I’d probably recommend switching to AWS - maintaining a 3 server deployment isn’t hard if you don’t need scaling, their support package is $30 per month, and we see pretty rapid responses to any queries we’ve had, though of course they can’t give any advice on Meteor itself.

I’ve not really used them for any Meteor advice so that’s a no-brainer.

Obviously, there’s also the cost of monitoring on AWS and I really like the Meteor APM package, but if I need to have an unreliable service to run with APM then that’s not really an option either.

1 Like

for monitoring we host our own Kadira server - its relatively easy to setup and we haven’t had to look at it since we started it, more than a year ago.

Thanks. That’s really interesting to know. I tend to use Datadog for most of my needs but with regards to Meteor it’s not been as nice as the APM offering which I believe is based on Kadira anyway.

We looked at datadig too. Because we host other services (mongo, pure node, etc) and kadira doesn’t catch everything. But kadiras method and pubsub stats and error traces are too good to pass on

DM me if you’d like any details on configuring AWS. It’s a bit of a hastle to get going, but once it’s setup it’s pretty nice

Dear Meteor Team,

this is now really moving towards a ridiculous situation. You have single-handedly without any advance notification suspended my account this morning UTC time (at around 9PM PST time) so after your office hours.

I had originally posted this expecting that you would respond promptly, that this would be seen by someone who might by chance have a look over the forums at around midnight of your local time and that it would result in a swift resolution and allow me to update my original post with a positive slant.

It is now nearly 11 AM on a Thursday morning at your offices, my account has been suspended for over 14 hours. Yet I have not

  • … had a response from Meteor Support
  • … seen any Meteor staff member reach out here to advise me to message them directly.
  • … had a response from Meteor Billing Support.

Whilst I don’t have any of your USD399 per annum support options with a 1 hour SLA, I have paid well over USD 1000 for your hosting services over the last 12 months. Yet I have not had any feedback whatsoever.

Can we just get this resolved as quickly as possible, please?

1 Like

I recommend heroku. Single command deployments, easy interface, no devops and quite cheap. I pay 7 dollar/month for my meteor app and it would be 25 dollar for a bigger multi az one.

@cloudspider I’m interested alternatives since Meteor can susbend accounts, so what would be the easiest approach to deploy to AWS perhaps with a single command ?

@hanskohls what has happened? I’m running 10 instances now on Galaxy. Now i’m afraid that they perhaps will susbend my account too. Wtf is going on?

Well, AWS is an IAAS. It’ will require multiple steps to do and things break rather quickly. This is why services like Meteor Galaxy and Heroku are so attractive. You pay for them to manage that.

The easiest way to manage something in AWS is to use something like AWS CodeStar, CircleCI or Bitbucket Pipelines. You can use those services to build your Meteor app into a node app and then deploy it. It only requires a commit to the right branch of your repository.

$1000? That’s merely the price of electricity and servers, not of people’s time. You’re having some unrealistic expectations here about default SLA levels.

How the Heroku deployments work? Any tutorial / examples?

Yes I know of some:

1 Like

This is so awful. I’m assuming that since no new posts were made, you’re actually back up and running. But really there’s only two options:

  • You aren’t up and meteor isn’t posting anything to not bring more attention to it.
  • You are up and meteor isn’t posting because they still don’t know how to do PR.

Either way, yikes.

Yikes - I’m sorry this happened @hanskohls! I know this doesn’t help much after the fact, but if something like this happens to anyone in the future, please ping myself, @abernix and/or @benjamn here in the forum. Filing a support ticket should definitely happen first, but if you haven’t received a timely response (and the situation is critical), please let us know. We don’t (usually) work directly on Galaxy support, but we can definitely help get visibility on issues like this (and help out with the issue backlog if the support team is swamped).

6 Likes