GCE plans are much easier to understand though. TIL AWS support plan cost varies by how much you bill and higher the bill, higher the support cost!. That doesn't make any sense.
Didn't they already tell you how to "architect applications for resilience"?
I was told that on more than one occasion. Essentially that one should be prepared if features don't work as advertised or unexpected things (including hardware issues) happen.
> First thought after 5 years of paying them a lot of money to migrate somewhere else (e.g. Google).
The thing that works in their favor is that some of the features tend to be a specific to the provider in question. The more one uses those features, chances are that they'd find it increasingly difficult to migrate elsewhere.
BTW – we always designed everything that there is no heavy lock-in on any provider, I would recommend that anybody else as well!
I haven't found anyone that loves bidding with Spot, and while I respect the ClusterK guys for the improvements they've brought to it, my hope is we convince them to move towards a fixed price as well. It's just simpler for customers.
Disclosure: I work on Google Cloud, and launched our Preemptible VMs product.
[0] https://cloudplatform.googleblog.com/2016/08/Preemptible-VMs...
Sounds like you had blind trust on your "partner" instead of building connection and paying at least for business support.
We're facing an auto scaling spot instance bug (it's definitely one) and we're trying since 3 days to contact anybody from them to get our business back on to the road!
We're now forced to sign-up for a paid support plan. Nevertheless, they already breached the SLA of 12 hours, it's really frustrating...
First thought after 5 years of paying them a lot of money to migrate somewhere else (e.g. Google).
It's always ciritcal if you lose your customer contact by implementing strange support barriers to earn 3$ more.
A not anymore happy AWS customer