Wholly agree. Too often we think about scale far too early.
I've seen very simple services get bogged down in needing to be "scalable" so they're built so they can be spun up or torn down easily. Then a load balancer is needed. Then an orchestration layer is needed so let's add Kubernetes. Then a shared state cache is needed so let's deploy Redis. Then we need some sort of networking layer so let's add a VPC. That's hard to configure though so let's infra-as-code it with terraform. Then wow that's a lot of infrastructure so let's hire an SRE team.
Now nobody is incentivized to remove said infrastructure because now jobs rely on it existing so it's ossified in the organization.
And that's how you end up with a simple web server that suddenly exploded into costing millions a year.
In a former job, I wrote a static PWA to do initial provisioning for robots. A tech would load the page, generate a QR code, and put it in front of the camera to program the robot.
When I looked into having this static page hosted on internal infra, it would have also needed minimum two dedicated oncalls, terraform, LB, containerization, security reviews, SLAs, etc.
I gave up after the second planning meeting and put it on my $5 VPS with a letsencrypt cert. That static page is still running today, having outlived not only the production line, but also the entire company.
Having a single process web/app server simplifies things operationally. I am building https://github.com/claceio/clace, which is an application server for teams to deploy internal tools. It runs as a single process, which implements the webserver (TLS certs management, request routing, OAuth etc) as well as an app server for deploying apps developed in any language (managing container lifecycle, app upgrades through GitOps etc).
Also, I think people vastly overestimate how much uptime their application really needs and vastly underestimate how reliable a single VPS can be.
I currently have VPSes running on both lowend and big cloud providers that have been running for years with no downtime except when it restarts for updates.
Shed a tear for Heroku, they made all this go away such a long time ago but ultimately squandered their innovation and the ~decade lead they had on other thinking in this fashion.
Could you please clarify? I haven't noticed any impact to Heroku on my web applications; it.. just works, anecdotally. They send periodic mandatory upgrade emails re database and application stack, but they have been harmless so far; going back a decade.
They went from leading / pioneering horizontal scalability and database deploying and scaling and orchestration to "quiet-quitting" 15 years ago and doing almost nothing ever since - today they're barely worthy of mention in any discussion on any tech that solves these problems.
I guess it depends on what world you live in. For example, using ASPNET Core, I just drop in this https://learn.microsoft.com/en-us/aspnet/core/performance/ra... and boom I have rate limiting and I do not have to stress about threads or state or whatever.
ASP.NET Core truly is a joy to work with
It's easy to have two implementations of your rate-limiting thing, in-process and Redis. Or change your implementation when you need it. Just put a nice interface in front of it.
There is a cost to the network synchronisation, so you definitely want to scale vertically until you really must scale horizontally.
The filmmakers of “A Quiet Place” or the “Terminator” franchises should employ cks as a technical consultant
and no race conditions whatsoever
Well, unless you're using threads.
Or did I miss the sarcasm?
Single-thread could be a thing if it's like a full-stack all sitting in a web browser--like Dioxus is going toward..
If a web browser is in a glorified chromebook like a 2025 Macbook Air, indeed there's a lot of breathing room. A lot of ram. Processing power. Cores. It's nice. I get that.
And then you can do off-line first: meaning use the cached local storage available to WASM apps.
Then whatever needs to go to the mother ship, then call web apis in the cloud.
That would, in theory, basically giving power back from "net pc theory of things" back to "fat client"--if you ask the grey-haired nerds among you. And you would gain something.
But outside of a glorified chromebook like a 2025 Macbook Air--we have to remember that we are working with all kinds of web devices--everything from crap phones to satellite servers with terabytes of ram--so the scalability story as we have it isn't entirely wrong.
I have been to U of Toronto, very smart people. But honestly this is a troll piece. Doesn't go into any depth and one-sided. Unhelpful. I think U of Toronto's reputation would be better served by something more sophisticated than this asinine blog entry.