Hacker News Reader

Comment by latch | Hacker News Reader

Rackspace is a tarnished brand for me. I immediately associate it with ridiculous pricing and bad "cloud" (mosso). With closed eyes it refused to see the quicksand of commodity it built its tower on.

I was hoping to come up with ways that companies with poor brands can recapture a lost audience. But I'm drawing a blank because it's a competitive market. It's like the car industry an the generational memory people have about various manufacturers. So, the only advice I have is: don't fuck with your brand.

View on HN

[Full disclosure: I am the PM for this project, a principal eng/manager at rackspace, and I believe in honesty]

> Rackspace is a tarnished brand for me. I immediately associate it with ridiculous pricing and bad "cloud" (mosso). With closed eyes it refused to see the quicksand of commodity it built its tower on.

You know what? I agree. Several years (4+) ago when RSC (Rackspace Cloud) came out I used it. My job was breaking the limits of cloud providers, and I did my job quite well. I was unimpressed with the cost and what I saw.

But - as old-gregg said much more eloquently; things have changed. From OnMetal to OpenStack, we've improved in capabilities, understanding the developer and hacker markets (my job, basically) and you're actually probably representative of the precise group of users I built this program for because this type of feedback, raw, unfiltered about technical capabilities and failings is what we need.

As someone on the inside now, I am proud to see the changes - the Managed Cloud launch was all about finally making the public statement "here are the infrastructure prices and here is the support you get". It's a clarifying move but a fundamental shift from the bundled pricing you had several weeks ago.

For example, when I went to developer events, people would ask me "do you have a cloud" - Yes I would say, and it has all these great features and capabilities and heck, we can do crazy topologies that mix dedicated gear with cloud stuff. People would be excited...

Then pricing would come up: and I would have to write down the prices and then explain that those prices included support - from 24x7 phone, to everything else you can imagine. But that triggers the uncanny valley when talking to developers as you probably know.

So now, here we are: we've separated our the cost of support from infrastructure. Our infrastructure prices are competitive with other providers (we lowered compute pricing for example), and now I have a channel with developer+ to try to win back - or get that raw feedback I need to make our products much better.

developer+ isn't evil, it's not got crazy ulterior motives: it is us saying developers are important, and we haven't done well for them. I want you to sign up, I want you to try the services, SDKs, Docs, dev portal for a spin. If it sucks, or there are bugs, I want to know directly from you.

In my mind, rebuilding a brand is much like rebuilding a relationship. It takes time, energy, honesty, candidness to regain trust.

And I'm saying we're here to do that. I can't promise everything is perfect, or we're not without flaws. I can't say developer+ is 100% to where I want it to be. But I'm aiming for that.

And I would honestly welcome you and your feedback, no matter how tough that love is.

My company hosted with Rackspace Cloud for nearly 3 years ($1k/mo average) but recently switched to AWS. Too much frustration and technical issues.

My biggest suggestion? Drop your 24/7 support or hire competent people. We had to contact support numerous times (a point I'll get to in a minute) and the experience was infuriating each and every time. It was the equivalent of talking to Comcast's tier-1 tech support. You guys would have blatant network outages and your "fanatical" support would be asking if I could ping servers (I had to point out the status page presence to them). We always had to waste time dealing with your ticket-jockeys before we'd get escalated and someone would actually solve the issue. I'd rather wait on a response than deal with an incompetent person asking for ping logs or suggesting I reboot a server.

Your private networking at ORD was problematic. We had our internal subnet at ORD which would very often just straight up fail. Server A could route to B, but the reverse was not possible. It appeared to be some sort of ARP fuck up, I dunno. We ran into issues every couple weeks and most often would resolve itself by the time we actually got a tech response. Which of course was closed with an equivalent of "could not reproduce"

Your MySQL-as-a-service is abysmal. It's running 5.1 (Do you have any idea how old that is?) and the performance is beyond horrible, I don't think the config was tuned whatsoever. Cloudfiles we routinely broke due to our write concurrency being too high (we were told you use sqlite?). This caused us to simply to manage our own software and completely avoid your pre-packaged SaaS offerings.

Since we switched to AWS we couldn't be happier and with reserved instance pricing we're actually saving money (annualized) on a superior platform.

Thank you. I'll actually send this directly to product management. All of this is valid feedback. As an aside, the MySQL offering just revamped quite a bit recently:

http://www.rackspace.com/blog/cloud-databases-introduces-enh...

So yes: we're acting on the limitations.

> developer+ isn't evil, it's not got crazy ulterior motives: it is us saying developers are important, and we haven't done well for them. I want you to sign up, I want you to try the services, SDKs, Docs, dev portal for a spin. If it sucks, or there are bugs, I want to know directly from you.

It's a marketing tactic. Unavailable to existing customers? Pure marketing.

"devs are important unless they already have an account"

> Rackspace is a tarnished brand for me.

I have dozen of servers with them (both cloud and hardware) and in terms of support and network they are pretty good.

If you used them back in the day, give it another shot. Totally different picture now.

> But I'm drawing a blank because it's a competitive market.

I don't think there is much competition when it comes to hosting mission critical applications.

Typical options are AWS or Rackspace. If you require any sort of dedicated hardware Rackspace is the only solid choice.

I tried putting stuff on DO last month for another start-up and so far their NY-EAST went down more than any of our cloud instances within Rackspace.

There are thousands (tens of thousands?) of dedicated providers, not to mention the possibility of colocating.

But here's the kicker, even if Rackspace was better than all the other ones and you were ok with paying the hefty price, in this not-so-new commodity reality, reliability and scalability is mostly a software issue, not an infrastructure issue. A lot of it comes down to getting a reliable messaging infrastructure in place (trivial) and writing software code that lives and breaths asynchronous messaging (not so trivial).

So is Rackspace the only choice? Not even by your definition. Softlayer's a clear alternative, for example. But beyond that, reliability goes way beyond the provider, it's 96% your code.

We are not talking about colocation/dedicated hardware. We are talking about cloud infrastructure.

The provider pool that offers a mix of both (and has proper integration between cloud and hardware network) is pretty limited.

Their cloud offerings (without managed support) are pretty reasonable priced in terms of $/bang. This is my opinion.

> reliability and scalability is mostly a software issue

Your software is not going to save you if your network is cheaply build and maintained. Your auto-scaling setup will not do anything if your provider is having network degradation issues.

I never said the Rackspace is the only choice, but it's one of the better choices when you look at the big picture.

Especially if you require actual hardware to be part of your infrastructure.

I have been Softlayer customer for 4 years (when they first started, during EV1 days). Their network and support is superb. I also recommend them to people who want to take the hardware route.

I did not mention them in my post because I have never played around with their cloud offerings.

They were a bit late to the game and I have already invested into integration with other cloud providers (API, etc).

Cheers.

Just on the softlayer-esque note, I'd love your thoughts about the OnMetal product

Have not tried it yet. If you do, please let me know about your experience!

I'm biased: I work for rackspace and fucking love OnMetal - I think it's the near perfect blend of API driven booting of bare metal servers I've seen. Last time I booted a High I/O instance it came up in 2.5 minutes. And the price / performance is a sweet spot.

Nice.

Do you have any pictures of OnMetal infrastructure? What do you guys have between the servers and the API layer?

> A lot of it comes down to getting a reliable messaging infrastructure in place (trivial).

Can you expand on this with some examples? We're currently spending a ton of time getting a RabbitMQ-based message infrastructure to work reliably in EC2 and I wouldn't really call it trivial.

I meant this in the sense that, for most cases, a typical deploy of something (most likely RabbitMQ) is going to meet their needs. There's a large community, stable libraries for every library and plenty of blog posts and tutorials.

Maybe you're doing something I consider atypical..maybe you're pushing more messages, have lower latency requirements, than I'm used to. But at that point, nothing "cloud" is going to help (at least, certainly not something like SQS).

For what it's worth, we essentially treat a write to our queue as fundamentally important as a write to our DB (wrapping the latter in a transaction of the former). This helps make sure that reliability issues shouldn't go unnoticed (I've personally never seen RabbitMQ swallowing messages, though it is something I've heard about, but seemingly always from relatively outdated versions).

You really should take the time to learn how to build your own server cluster. The savings is exponential. The control is complete. The knowledge is priceless.

> You really should take the time to learn how to build your own server cluster.

Not sure where I mentioned that I don't how to do that. But thanks?

> The savings is exponential.

For a lot of start-ups rolling their own infrastructure is generally a giant waste of time and money in both short and long term run.

>For a lot of start-ups rolling their own infrastructure is generally a giant waste of time and money in both short and long term run.

Exactly. Which is why people go to Rackspace: not having to hire their own in-house DevOps guy (which doesn't scale as your servers grow) to take care of their servers.

Savings? What things do you add up, or forget to add up, to come to that conclusion?

The control is complete, as is the obliteration of your infrastructure and/or your data. Because let's face it, unless you invest the same millions(in money and hours) as Rackspace(or any equivalent company), the odds are against you :)

The knowledge is indeed priceless, if you're into that sort of thing. Some people just need to host their application... reliably.

Sorry but that's like saying building a nuclear power plant to power your christmas tree in your apartment is a good idea. I agree it can be nice to learn these things but in a mission critical environment you should focus your time on what adds real value to your customers. Unless you're building the next AWS or Rackspace doing these things yourself and saving $50.00 a month is not the right place to focus your efforts.

Part of the problem is that the responsibility is also complete. And the management overhead at small scale can be pretty significant. Redundant power/networking, backups, etc.

Thats a pretty bold claim that Rackspace is the only solid choice. There are many other providers out there that provide better white glove service.

Can you (honest question) itemize what's better white glove service? Because I want to take that list and act on it inside of Rackspace.

The only other 'white-glove' provider I've used was Voxrox (Which became Voxel, and then got bought out by Internap iirc) - They had amazing support up until they got bought out by Internap. It really boils down to getting extremely competent support on the first try every time and having 911's answered within seconds. It's a pretty amazing thing to be woken up at 2AM by your own monitoring, put in a 911 ticket, and literally get a actionable response in seconds.

The place I am at now has servers at Rackspace and we can't even get ticketing and other alerts to go through systems like PagerDuty from your own internal systems.

Ok; so since I'm in product manager mode:

* 911 style issues answered in seconds over phone or chat * Integration with our ticketing system with PagerDuty / other tools.

Question: Cloud Monitoring is now free for all customers - AFAIK (we use it) CM integrates well with PagerDuty (which cough I use a lot). Are you looking for ticketing responses into the PD API, or all monitoring/etc?

The product manager below clarified that it's a 1 year free trial... so you're pretty spot on.

EDIT: I'm just realizing now that the product manager is Jesse Noller, prolific Python and PyCon contributor... so maybe it's not evil at the root after all.

It's really not evil. Seriously

Same. They rapidly went from one of my favourite companies in the world (they were unbelievable in the late nineties into the early 2000s) to a total farce. Today we still use them as a secondary, mainly b/c we're working on other things, and their London data centre is beyond horrible, their customer service nonexistent, and everything is slow, expensive, and always breaking compared to others.

Reply on news.ycombinator.com