It can also write good kubernetes GitOps files, although that doesnt necessarily make terraform or ansible unnecessary (especially if you have to set up a diverse set of resources at a cloud provider)
Do they still maintain it as in, it gets bug fixes and security patches? Technically yes, bug fixes and security patches are still produced for Windows Server. I wouldn't call it a good level of effort but it exists. They are happy to take your money for such a mediocre level of effort though.
Is it Microsoft happy path? No. Microsoft STRONGLY recommends that if you are running .Net (Core), you run under Kestrel on Linux in a container. Performance under Linux is much better than Windows.
Do any web servers have meaningful updates? It’s pretty standard stuff. IIS can do quite a bit of the same things. It’s common in my industry. We set up URL rewrites, reverse proxys, asp.net, even node.js or python (flask) based apps using WSGI. Separate app pools for everything all on the same server sometimes. It’s never been a bottleneck because it’s pretty streamlined code. Haven't had any security issues with the web server ever. It’s nice there is a GUI but when you start looking at the configuration (xml based) a lot of the directives are familiar if you’re used to nginx.conf or httpd and with htaccess (web.config) in site directories.
It has newest TLS stuff like TLS 1.3, it has HTTP2/HTTP3 and QUICK implemented. I do think those are meaningful updates, well http2 was 11 years ago but QUICK is more like 3 years ago with HTTP3.
Correct me if I am wrong but I always thought http.sys is IIS.
I guess if they can do changes to SCHANNEL but not to IIS for TLS stuff that would be good engineering. But it also makes argument mute ... because if Apache and Nginx and Caddy get updates to support new TLS and IIS is not because it is separate, then a question who does it right?
Sharepoint, Exchange webmail, and other stuff uses it... that is why it's officially NOT depricated... but man imagine having to keep up an ASP codebase today, it's not unheard of, it's just rough.
I'm kind of having a hard time understanding what problem this is trying to solve? I can't even tell what this is actually doing from the site. There is barely any information.
I mean Visual Studio automatically has a way of deploying directly to IIS for years now. Also as others have added, IIS isn't even used anymore with the rise of caddy and nginx.
The pricing itself seems really steep as well $3K a year?!? I don't exactly know who this is targeting.
No one connects to prod directly from VS in corporate setup. You do it via pipelines.
It is for when you want to have accountability of deployments and at least what we had in Octopus deploy I could promote versions from test/acc to prod based on triggers or manual click of a button.
So you can have build pipelines separate from actual deploy pipelines and manage artefacts built by the build pipeline. It is required in some environments, but totally useless and waste of time in smaller environments.
Like the sibling comment mentioned, there's still lots of .net framework apps that target IIS. I haven't used IIS in about 15 years, but this looks like it provides a similar feature set to what Octopus Deploy had back then (Octopus seems to have changed a lot since then). i.e. more than what VS deployment can do.
There is still plenty of Window Server using enterprises churning out .Net Framework code that must run on IIS. It's shrinking market for sure but it's still a market.
If they are willing to spend money on Windows Server, 3k a year to make it not less painful is probably worth it. Remember, any company using Windows Server is not hiring skilled DevOps/SRE/Ops, it's probably average at best, low paid workers. It's the type of people who have probably never heard of HackerNews either so it makes advertising here an interesting choice.
we still use windows server and IIS, and I can assure you we're never spending 3k to make anything less painful. our middle name is pain. and we bill hourly for that
There probably are businesses doing clickops who'd spend $2900 on a tool, but that market is rapidly shrinking. AI can write pretty decent terraform.
1. well done.
2. I don’t tell you my troubles.
Is it Microsoft happy path? No. Microsoft STRONGLY recommends that if you are running .Net (Core), you run under Kestrel on Linux in a container. Performance under Linux is much better than Windows.
https://aka.ms/aspnet/benchmarks
Here is Microsoft official ones, Linux beats out Windows in almost every metric.
I guess if they can do changes to SCHANNEL but not to IIS for TLS stuff that would be good engineering. But it also makes argument mute ... because if Apache and Nginx and Caddy get updates to support new TLS and IIS is not because it is separate, then a question who does it right?
I mean Visual Studio automatically has a way of deploying directly to IIS for years now. Also as others have added, IIS isn't even used anymore with the rise of caddy and nginx.
The pricing itself seems really steep as well $3K a year?!? I don't exactly know who this is targeting.
It is for when you want to have accountability of deployments and at least what we had in Octopus deploy I could promote versions from test/acc to prod based on triggers or manual click of a button.
So you can have build pipelines separate from actual deploy pipelines and manage artefacts built by the build pipeline. It is required in some environments, but totally useless and waste of time in smaller environments.
If they are willing to spend money on Windows Server, 3k a year to make it not less painful is probably worth it. Remember, any company using Windows Server is not hiring skilled DevOps/SRE/Ops, it's probably average at best, low paid workers. It's the type of people who have probably never heard of HackerNews either so it makes advertising here an interesting choice.
Not painful enough to justify additional tooling in any case.