Hi there and welcome to Scale. I am your host, Stewart Ritchie, founder and CEO/lead developer at Power by Coffee. Skill is a podcast about media and technology, how technology impacts the media and how media is, in turn, impacted by technology. Power by Coffee is a open source web development and software development agency focused on media, solving problems as and when they come up. Today we have a guest, david Smith. David is a lead DevOps engineer with Pagely. Pagely are a leading provider of managed hosting services for WordPress, and their clients include many, many media companies at various sizes and scales and operations, some very small and then some of the very biggest media conglomerates in the world. David, first of all, thank you very much for being on.
And thank you for having me, sir, no problem.
I want to start with the job title DevOps Engineer. I think DevOps is the thing that’s kind of cropped up in the last few years, but I think a lot of our listenership is probably outside of the technology developer world, listening to kind of understand some jargon that they hear and some terms. So I really think it’d be a great idea for us to just start with DevOps. What is DevOps and what does it DevOps engineer to?
DevOps is one of those words that means exactly what you want it to mean. It means whatever HR has decided it means this week. Pagely is a pretty small company, so for for me it means a little bit of everything. Pagely has built our own hosting stack and we’ve developed some custom software that goes along with that. Plus, we work on all of the server automation so that when new customers come in they get they get all of their servers, everything they need set up right away, all the integration to tie our custom software to a third party billing platform, to a third party help desk. It means for us it means a little bit of everything.
Well, it’s really the glue, that’s kind of sticking all the pieces together.
Yeah, that’s a great way to put it We are glue. We are holding all the other little pieces of the internet together.
Awesome. So you’ve got, for example, like you say, your third party kind of like support desk in your third party billing and you’ve got the admin panel that’s kind of proprietary ish to Pagely, how that all connects together And then within that how all of that ties to how the actual servers are managed. So the tools that you’re running to add new hardware to the fleet and then install a new user or new customers set of sites and all the bits and pieces that that they need to do that, is that about right?
Awesome And I suppose with that it’s really important that that be absolutely repeatable and absolutely the same every time it runs, And there could be no real for variation within that.
Exactly, there are none on the scale that we operate. It’s very important, and that is one of our guiding principles. If we do something custom for a customer and we do that pretty frequently actually we make sure that everything is documented and that everything gets put into code. We mainly use a tool called Ansible to help automate deploying code and software and configuration changes out to the fleet And we make sure that everything is saved there. In a sense, the code is the business.
Absolutely Cool. Well, that’s great. So now that we’ve kind of established that baseline of like what is DevOps, let’s kind of come back to kind of what we normally go through. David, tell us about yourself, Tell us about Pagely, tell you how you got, got where you are And what it, what it looks like at Pagely for you.
Yeah Well, as I mentioned before, for me a day is a little bit of everything helping respond to customer tickets and helping develop new tools and automations to make things work even even more smoothly than they already do Keeping the trains running, as they say. I have been so. I started working at a very small little mom and pop internet service provider in the Midwest United States back when dial up modems were a thing And website hosting was just sort of a sideline, and from there I moved on. I spent several years working in higher education for Washington University in St Louis. Essentially, they hired me to be the WordPress guy and the Linux guy.
And plenty of audience knew it, knew it well.
Exactly, and from there I had a great opportunity to move on to Pagely, to get to spend even more time in the hosting space, which was always my favorite part of working for that smaller ISP back in the day. And here we are today.
Absolutely So, then, and tell us about Pagely more explicitly, i suppose, because there are hundreds of thousands of web hosts out there. So what? what kind of sets Pagely apart? I mean, obviously there’s the managed WordPress site, which is kind of the crux of it, but even if you didn’t think WordPress and just like you know running a PHP application on this, what makes, what makes Pagely different from everyone else?
Pagely’s focus is what I like to think of as upscale WordPress hosting. You’re absolutely right in that. There are lots of places that will give you a shared account on a shared server for your little personal blog and your mom’s recipes for a few dollars a month, for not a lot of money. And if you’re just hosting something very small like that, honestly that’s probably fine. But if this is something where it’s a big part of your business, if it is your business, then you’re going to want something a bit more upscale. You’re going to want dedicated servers. You’re going to want dedicated resources for your website and for your database. You’re going to want experts that are on call for you 24 hours a day that can help if you do have problems, and that really is where Pagely comes in.
Absolutely. There are also many, many hosts now that focus on WordPress. Obviously, page has been around for a long time. We say 2009, certainly one of the first I recall coming across as like a. We are a WordPress host rather than a. Here is a VPS you do what you want, and here is, you know, a shared hosting account. So definitely in that first kind of rind of managed hosting providers. But what do you? what is kind of Page’s differentiator between the likes of, you say, kinsta or WP Engine or some of the other players in that space?
Sure, obviously, we have been around longer than most of those other companies We have. We have focused entirely on building our infrastructure within Amazon Web Services, within the AWS cloud, because they have been around longer than anybody in the public cloud space They’ve done it the longest and we think that they do it the best and we have built our own software on top of that. We have our own custom management platform, we have our own custom server-fied software and everything is designed with the WordPress and the AWS integrations top of mind, just to so that you, the customer, have everything put together in in one platform, in one place.
All right. Okay, for my experience, aws is incredibly powerful but an incredibly complex tool. But you don’t see a lot of that exposed you within Pagely. I have found it hides it really well by how complex like that kind of top tier cloud hosting really is.
Yeah, absolutely. For most businesses, we actually would probably recommend starting on a plan that, coincidentally, we call our performance plan. All right, and yeah, thanks for the plug. And that gets you into the Pagely platform, gets you to Pagely experience at a relatively low cost. You will get a dedicated server. You’ll get access to all of our expertise. We’d be glad to help you out, depending on what kind of content you’re working. If, say, you’re a smaller podcaster working on distributing your podcast to the world, we might be able to help you integrate into the Amazon, into the AWS CloudFront content distribution network so that your podcasts are globally available with the best possible latency, so that people streaming or downloading your podcast will have the leaf amount of content buffering and hiccups.
Absolutely So. I mean, one of the things that these publishers we have always find are the things they tend to be most concerned about, it’s how performant the sites are. So everybody’s really really very concerned all the time about their core web vitals as an example. And one of those core web vitals is time to first bite is what I’ve always called it. That’s been renamed within core web vitals to something else. That’s kind of part of the bigger one, but it’s like how long does the server tick to respond to this request?
Oh, there are. It’s caches all the way down And that is another big part of PageLeave. Secret soft is that we have as part of our server-side platform. We have content caching. So the front page of your site, which I imagine for most websites is going to be where most of your visitors start That will already be cached as soon as someone visits it. We generate that content and save it for however long you need it.
Absolutely So. Is this perhaps more closely aligned to the CDN slash content delivery network more than the Press 3 service that is offloading your kind of static assets into S3? Presumably, so that the CDN can then Amazon’s CloudFront it’s CloudFront, not CloudFlare. Always get the two confused.
So that’s, a lot.
So that you are, and that’s not like an add-on, that’s not something that we’re having to do inside of WordPress. That is part of the hosting. It is kind of handled automatically by the platform.
It’s not enabled by default because it does add a little bit of extra complexity. So we generally recommend that you get the website on-boarded first to make sure that everything works, and then you can go to the control panel and you can enable some of these extra features. You can contact support to get Press 3. You can enable the Press CDN, which is our name for integrating into CloudFront, and that’s just a couple of buttons within our management interface. You start with the basics and then you can build on from there. You can add on the extra complexity and the extra performance.
And of course that is one of the benefits of PageLives that we’ve done all this hard work for you. Setting up CloudFront if the polite way to say it would be inconvenient, awkward pain in the behind maybe. Yeah, that’s fair, and we’ve done all of that hard work for you.
Yes. So even then, at the very kind of like most, at the kind of kicking off, like you’re me to say you’re just getting started. You know good, manage hosting, pagelive and many others. You’re looking at that CDN integration into CloudFront. In this case, you’re not having to take care of, you’re not having to configure it yourself. It’s just part of the platform, part of the hosting And, additionally, that layer of on-server caching for page generation, so that there isn’t a need to be installing with, frankly, very fragile and unreliable WordPress plugins that make this very difficult to do and difficult to manage, is just part of how the hosting platform works. Is that about right?
Exactly Those third-party plugins. They do exist and they do work well enough, but most of them are something that you have to pay extra for And They are third-party plug-ins. So the web request has to go all the way through the stack, through the web server down to the actual WordPress software, before it can pull this cached content back out. And by having caching be a part of the Pagely platform, that just saves a few more cycles. It means that that page load is just another few milliseconds even faster, and it’s one less thing that you, as site owner, have to think about.
It sounds awesome. It is awesome because we’ve used it. So then, i think that’s a great starting place for someone who’s coming in to look at the hosting of where their site is going to live. Is it going to be fast? Is it going to be performing? Is it going to stay fast under traffic and under load? But what about now for that next group up?
Sure, We do offer CI-CD tooling, continuous integration, continuous delivery for people who love their acronyms, but I think everyone just calls it CI-CD which is essentially a way to hook your source code into the web server so that when you make a change and you commit this change to your source code repository, you can have it automatically get deployed to the web server and that new code is there within moment, so that whenever you change something, it’s ready to go and you can make sure that everything is still working. You can validate those changes immediately. We hook into all of the major providers. We have integrations for GitHub probably our preferred one but we also have integrations to GitLab and Bitbucket server and one or two other providers whose names escape me, but if it’s somebody that we haven’t worked with, we’d still love to hear from you. We can probably help you out with it anyways, over just the ones that we have the most experience with.
Okay, and that’s very much within the developer tooling realm there. So, developer, whoever’s working on the site generally, will have all the source code in a repository on kind of, like you say, one of those services, and effectively there are rules that can be set up. So when a change is made in one of those repositories, a process is kicked off that moves all that source code, compiles it, does the actions it needs to get it into the shape of a website and then replaces the existing deployed website within page.
And you can and you can and probably should have all of this go through a couple of levels first. It’s very common to set up a test or staging website and you push everything to that site first to make sure that the code doesn’t have oh, I overlooked a semi-colon and now my entire site has crashed, which happens more often than you might think. Yeah, And we also have tooling. you know, if you have recently made a bunch of big changes to your production site, you’ve just added a bunch of new content. We also have tooling to help you sync all of that content from your production site down to your test and staging sites so that you have the most current content available there, to make sure that new podcasts you just put out doesn’t somehow cause problems with your website’s theme or anything silly like that.
Absolutely Security. One is a very interesting one because there are so many different ways that you know this can go like. There’s obviously the security of kind of the user, kinds of like people who have access to Pagely’s control panel, you know you might have, you know, technical leadership within the company as individual developers, billing, you know all those kind of different accounts that can get access into the control panel. Whole nother layer of security of, like well, which of these users should actually be able to access the server and, crucially, which parts of the server should these users be able to access? like this developer can only access, for example, this site and this site for you know, variety of reasons internal to that company And that’s beyond the kind of general like the internet is very hostile to everything. But yeah, does that make sense?
Yeah, yes, And Pagely have taken a lot of those things into consideration already. to get into our control panel, Obviously there has to be one super admin who sets up the account and that has access to everything, But then you can create other contacts within your company that can only access the billing side of things. you can create users that have access to the servers but not the billing side, and have people who have only FSH and SFTP access to individual sites, but not all of them We can. we have a lot of flexibility in terms of making sure that access to that part of the business can be locked down. And there’s, of course, all sorts of other things, things. there’s access to the server itself, which is locked down to a subset of Pagely staff and those people that you specify that should be able to SFH into the server, And that is not enabled by default for everyone that you add to an account There is.
Absolutely, i know. One of the things we’ve also noticed is things like, particularly within the WordPress world WordPress itself is actually surprisingly secure. It’s the plug-ins and the ecosystem around it that generally so. Even notifications of there’s been a vulnerability find in X, y or Z plug-in. Automated scanning that we are running is telling us that you are running this version. Please ensure that it is updated quickly. Coming through as support tickets raised for us from Pagely.
Exactly, Depending on the vulnerability. if it’s something very severe, we may just go ahead and update it for you.
Yeah, absolutely. The last thing within that realm is the compliance piece. Obviously, compliance the legal frameworks of things you’re hosting and infrastructure have to meet can vary by jurisdiction, but is that something Pagely is able to help us with as well? One of the obvious ones is where is data located? where does it live? If we’re a European company, is it possible to have our data and the data centers we’re using be within Amazon’s European centers rather than the US, for example?
Of course That is. Another. One of the reasons that Pagely work with AF exclusively for our hosting needs is that we have data center on every continent except Antarctica, and I’m sure they’re working on that one somehow.
We can put all of your resources, your databases, your virtual servers in just about any location that they offer service.
Awesome Good stuff. If we’re looking at the entirety of the small and startup sites and these scaling up media brands within Pagely’s perspective, there’s no actual distinction there. All those tools are available to everyone, regardless of what they’re doing. It’s the number of server assets and how much resources you’re using up within Amazon that is the differentiator. Is that about right?
Correct. We use the same software on everything from the smallest low end, from the developer plans all the way up to through performance plans all the way up to the enterprise plans, all the way up to high availability plans that involve multiple servers scaled across different data centers and sometimes even different geographic regions. It’s the same software all the way down. It works pretty much identically, regardless of your environment.
Absolutely Well then. That then brings us, nicely, one of our last group of these companies, what I call platform media companies. Again, folks who have lots of brands, lots of different touch points within those brands and who can be handling characteristically huge traffic spikes, but not just a single traffic spike necessarily. They might have multiple viral articles or multiple viral pieces of content across their different brands all at the same time. How do we handle this within Pagely? You mentioned our scaling across multiple machines. Is that the answer here that we’re looking at?
Sometime. We usually recommend and there are good arguments both ways on this, but we usually recommend keeping as much as you can, keeping things on a single server and scaling that server up, because there’s a lot less complexity to it. If everything is on one web server, you have this single source of truth. This is where all of your media assets are. This is where all of your code is running. As soon as you scale that out all of a sudden you have a whole new stack of problems.
Is it more likely that someone needs that multi location, multiple server set up to handle traffic or to handle reliability. So, for example, like I say, aws really good, really reliable, but not infallible. So you maybe do get an availability zone goes down and that could take a lot down. You know a huge chunk of servers. So then the user can swap over to the other availability zone.
But probably the reliability case.
Yeah, you want to have, you want to have everything.
No, if you want your content to be available all the time, as close to 100% of its humanly possible to do, then you need that replication, you need that redundancy that we do have some customers that have something like that just because of the volume of traffic that they handle, but there’s not as many of those as there are the customers that need the reliability and the redundancy.
Awesome. Are there? do you ever find situations where I only ask is we’ve had this like happen once or twice where sites are relatively low traffic but they are using almost all of a EC2 or an Amazon compute unit that they are accessible on, but any the occasion they get spikes. This is where we have used things like auto scaling on AWS. It was like you need a spike, need to be able to handle it to drop back down, and that happens occasionally. Yeah, i don’t know where I was going with that, it’s just a good like thinking about it, yeah, and that is another thing that comes up.
If you know, if you’re an e-commerce site and you know that you’re having a big sale next week, reach out to us. we’d be glad to preemptively upscale your server for a few days so that it has those extra resources. We don’t generally do the kind of auto scaling that you described, because that’s even more complexity on top of the usual. If you have two web servers that are running all the time and you have to keep content in sync between them, that’s a lot of complexity. And if you have more servers than that that can come and go dynamically, that just adds it up more and more and more, and that just it’s more things that can go wrong And we don’t think that the benefit is worth the added complexity in that case.
Absolutely, it makes a lot of sense. David, i just know we’re coming towards the end of the time we have available. So, first off, thanks very much for going through all of that with us. I know that can be really difficult in some ways to think through kind of what the differences in what is appropriate level of hosting for some clients And thank you for your expertise and kind of what pagely And, to be honest, a lot of the other hosts, but we like pagely is to be able to provide to tailor to these situations. If you want to find out more about you, where can they go?
Sure, if you’re interested in me, you can go to davidsmithif. That’s davidsmith.is. And if you want to find out about pagely, we’d love to hear from you. We’re at pagely.com.
Pagely.com. Thank you very much, David And, if you are still listening, thank you very much for making it through. I would really appreciate if you could leave Scale a review wherever you get your podcast, whether that’s iTunes, the Google Play podcast app or wherever. If you would like to follow Scale, we are Scale Podcast on Twitter. There’s also going to be a link tree in the show notes that will take you to everywhere you can follow this. I’m Stewart Ritchie. Again, you can follow me on Twitter. I’m Masadon Link’s. All in the show notes. David, thank you again for your time. I really appreciate it. Thank you for having me.