Comment by DanielBMarkham | Hacker News Reader

Comment by DanielBMarkham | Hacker News Reader

Parent Full thread

I don't know, Patrick, it really seems like a lot of pain and suffering you went though for nothing.

Perhaps if you added your actual volume numbers to your entry it would help the conversation? I have several sites on various platforms, and I'm currently running over 160K visits per month across all of them. It's not unusual to have 10K visitors from HN drop by on any given day. I haven't had any problem with WP. Of course, I don't really check it. It just works.

I hate to sound non-caring, but is it really a problem if ten people come by your blog at 3am and can't load the site up? They'll be back at 4am, and the site will be fine then. Do you really need 99.999 uptime on a blog? It's all a numbers game. If you lose 2% of visitors due to technical problems it shouldn't be a huge deal. Just take a look at your funnel and conversion numbers and do the math. There's no way that damage is going to come out to $2500 per year.

On a personal note, what I find is that there are a huge number of opportunities for my startups to accrue costs that don't justify the benefit. It's tough figuring out what to spend money on. Things like hosting or tools have to pull their own weight.

ADD: There are a lot of companies that made huge amounts of money leveraging free blogging services. Not my cup of tea, but it does provide a bit of context.

3 AM Japan time is, um, what time is it for you now? It's then. Right in the middle of the US work day. Which is typically when Jimmy Wales decides to get me mentioned in the NYT. That was mostly not business-oriented, but in general, that sort of thing is very worthwhile.

I don't really have a conversion funnel for the blog at present. (That will probably change in the future, which is one of the reasons I knew I had to switch.) It's just my main source of consulting leads and a) my rates justify $2.5k pretty easily and b) consulting clients start getting all kinds of skeptical if their SEO guy isn't on the Googles because his server went down.

(I'm not the only person with this problem. You should have seen Fog Creek when their blog decided to die. By freak accident I was in the office when it happened and was able to successfully diagnose the problem from "Wordpress just crashed": it was Apache KeepAlive because it's always Apache KeepAlive, grumblegrumblegrumble.)

So this is really more of an image thing -- site responsiveness, accessibility, and so forth. Similar to how a shopkeeper might have paid 2500 each year to have a sign custom-painted out front.

In that case, makes sense. I'm not an extreme cheapo -- I'm probably paying 1800 a year for my various hosting solutions, but that's spread across HostGator, a MT hosting company, and AWS. I'm moving to AWS -- exactly in the opposite direction that you are heading -- because I like buying cycles. I can set up databases, write code, and have more power to write hybrid sites. I'm beginning to look at producing content on the web, blogs, apps, static sites, or mixed-mode stuff, as more of an art form and less of a pre-canned WP experience. So I need a box, not a service.

But I look at this as simple content. I'm serving bits to folks who come and consume them. I don't generate consulting leads though my sites, I provide data people want for free, so as long as most of the people get most of the data, everything is copasetic on my end. It sounds like we have completely different purposes.

Just as a off-the-wall suggestion to other readers, you know you could write a script to consume your blog entries and move them to a CDN as static pages. You'd get a hell of a boost in performance, and I don't think you'd really have to change much. (Probably much tougher than I make it out to be. Haven't really thought it through. I use URL rewriting on some of my newer sites, so it all looks like static content anyway.)

How is that an 'image thing'? He's a professional programmer and SEO consultant. It's not an 'image thing'. It's a basic 'this is my job, why isn't it working' thing. That it's still such a pita in this day and age that you need professional keeper-uppers because it's so opaque and complicated that web programmers have no real clue what we're supposed to do is all a bit odd.

Taking your analogy it's is more akin to the tills in the store keep crashing in the middle of the lunchtime rush.

You can walk around and tell you customers it's the till software suppliers, but in the end it just reflects badly on you. Or you can follow your suggestion and get out a pad, a pen and a calculator and write out their bill. Neither is a good solution.

You should only be creating custom solutions if this (blogging for instance) is your primary business. If you want to kick ass as a one-person team you have to know your limits and only spend time of things which are critical.

Nice suggestion re CDN.

You should only be creating custom solutions if this (blogging for instance) is your primary business

Even more specifically, only if existing tools do not meet or cannot be adapted to your needs!

I disagree. If something is core to your business you cannot outsource it.

Google couldn't use someone else's search engine

Microsoft couldn't use an outside kernel or filesystem for windows

Gridspy has to make its own dashboards. However, We can use handy libraries for html templating and graph rendering.

But a blogger (the given example) produces content. A blogging platform is core to their business, but it isn't what brings value. In Google's case, the search engine is what brings value.

Don't let yourself get caught up in NIH syndrome.

But look at it another way : Maybe your hands-on knowledge of KeepAlive is the kind of thing that signals a client that you're not just a regular SEO guy...

I think this could be a great route to take. Patrick might be at the point that he has enough marketing work to do though, that he wants to focus on that.

Many of us have $1000/day+ contracting rates. Beter to spend the extra week contracting per year rather than upgrading plugins and securing servers.

I solved the problem by using movable type which doesn't serve dynamic pages. Still heaps of maintence though.

Course: I serve no traffic.

Http://blog.gridspy.co.nz/

Our blog is hosted as part of our general website, and it lives in 4x servers at AWS. Did you ever consider load balancing across 2x512 instances rather than 1x1024mb server? We considered options like outsourcing the hosting of elements of the sales site, however its just part of our standard build and in reality i don't see any major issues with load balancing a WordPress blog.

Your main bottle neck once your WordPress blog is under load will become your DB, but if that's your problem you're probably doing something else very right (so congrats?)

Reply on news.ycombinator.com