3 ms·

Meh, as a developer who lived through the 2000-2010 era Microsoft, it’s easy to come up with a laundry list of reasons to hate on Azure. But I tried Azure for

Meh, as a developer who lived through the 2000-2010 era Microsoft, it’s easy to come up with a laundry list of reasons to hate on Azure. But I tried Azure for my most recent startup because I was offended by AWS, and GCP did not have enough adoption among my customers, and Azure worked - fine. What do you really need out of a cloud? I want them to rent me VMs, for them to not go down, and to make it easy to do standard stuff like an object store, run containers, run databases, etc. Azure was as good or better than AWS

deanCommie 2y ago
Why were you offended by AWS?
- gdiamos 2y ago
  I got into a fight with them kind of like Trump and Zelensky just did. Not a technical reason. Sometimes in business the deal falls through I was on the receiving end and didn’t appreciate it.
jml78 2y ago
I have used all the platforms personally and professionally. GCP, AWS, Azure, Oracle Cloud. I will just say that Azure seems to want to do shit different for the sake of being different. It is really annoying to write infrastructure as code for aws/gcp then go to do the same for azure and realize how dumb some of their stuff is. Just my personal experience.
- gdiamos 2y ago
  That’s how clouds try to lock you in, by making you use a custom tool that is different for the sake of being different. If you use standard tools you don’t have this problem. Containers running on VMs is standard. A mesh of microservices that depend on cloud queues and managed services is not. One argument against standard containers is saving dev time. You can still save dev time by using standard open source software. How many different ways are there to implement a queue or a load balancer? If you really need access to some proprietary technology then by all means use the cloud that offers it. Eg if your customer demands GPT4.5, then go with Azure. But if you need something standard, don’t get caught in the trap.
  - jml78 2y ago
    I am an older guy that was building kubernetes clusters before eks, aks, gke. So I used terraform to build shit out to make it happen. Azure was 5x the code just to be different. You can try to blame terraform but if you used MS custom tooling it was no different.
  - rconti 2y ago
    What about the way Terraform is a 3rd class citizen on Azure? And there are multiple ever-changing ways of doing everything, major parameters aren't supported, etc. It just makes it more difficult to deal with. Also, Azure APIs are incredibly slow.
    - re-thc 2y ago
      It’s Bicep
      - rconti 2y ago
        I'd rather avoid the trap. I use the 3 major CSPs, so I would prefer to use cross-platform tooling. Per the parent: >That’s how clouds try to lock you in, by making you use a custom tool that is different for the sake of being different. > If you use standard tools you don’t have this problem.
- jiggawatts 2y ago
  Conversely, doing things "the same way" as AWS would mean copying their first-generation public cloud design flaws. The overall UX of AWS is absolutely crazy. It's easy to "lose" a resource... in there... somewhere... in one of the many portals, in some region, costing you money! Meanwhile, Azure shows you a single pane of glass across all resource types in all regions. It's also fairly trivial to query across all subscriptions (equivalent to AWS accounts). Similarly, AWS insists on peppering their UI with random-looking internal identifiers. These are meaningless and not sortable in any useful way. Azure in comparison allows users to specify grouping by "english" resource group names and then resources within them also have user-specified names. The only random identifiers are the Subscription GUIDs, but even those have user-assignable display name aliases. The unified Portal and scripting experience of Azure Resource Manager is a true "second generation" public cloud, and is much closer to the experience of using Kubernetes, which is also a "second gen" system developed out of Borg. E.g.: In K8s you also get a single-pane-of glass, human-named namespaces (=resource groups), human-named workloads, etc...
  - smackeyacky 2y ago
    A single pane of glass that shows all your resources that are currently choking due to hidden limitations is no flex over AWS. It is my hope I never have to use Azure ever again professionally or otherwise
- iamtheworstdev 2y ago
  this killed me. i dont think AWS does everything well... but Azure went hard on being different just to be different.. like region names...
- stego-tech 2y ago
  > I will just say that Azure seems to want to do shit different for the sake of being different. That's Microsoft's MO in a nutshell in my experience, and I say this as a recent(~5yrs ago) convert to Linux who built a career on Windows endpoints, servers, ADDS, Exchange, SCCM, you name it. It's how they achieve lock-in to their ecosystem, and it's incredibly frustrating to see how they've just layered that method of operation over and over again, decade after decade, rather than fix anything.
  - Henchman21 2y ago
    Fixing things is hard. Papering thing over with free Azure credits and marketing? This is the way.
BoredPositron 2y ago
The UX not being designed by business majors.
- inetknght 2y ago
  You won't find that in major cloud providers.
zeroxfe 2y ago
> What do you really need out of a cloud? I need a cloud to be reliable and secure. I've used Azure extensively and it's neither. I'll take GCP or AWS over Azure any day.
- foxylad 2y ago
  This. We've used GCP Appengine for years and it is rock solid. Their SRE game is top level, and when there is an outage, they do a serious investigation and make it fully public, even if they screwed up badly. Including the vital "this is how we're going to stop this ever happening again". The last outage (that we noticed) was several years ago.
avereveard 2y ago
Azure tend to bite back when they upgrade their backend and everything breaks Happened twice with my kubernetes deportment, first something with node groups made them incompatible and had to recreate the cluster from scratch, then one of their scripts to rotate key access to volumes (that one has to run manually, go figure) stopped working and caused my volumes to detach from pods, and has to recreate the cluster again, and I just give up. I was super happy as well, the first two years. By years six I was fully migrated out.
FridgeSeal 2y ago
I need it to be sane, to work and to be reasonably well documented. Azure ostensibly fails outright on points 1 and 3, and limps by on 2. The products are confusing mess, there’s way too many ways to auth things, docs are garbage, tried multiple times to manage stuff via Terraform which broke far too much to be excusable, to say nothing of the dumpster fire that is their UX. I’m sure some people have either beaten it into submission, or have stockholmed themselves into putting it up with it.

deanCommie 2y ago

Why were you offended by AWS?

gdiamos 2y ago
I got into a fight with them kind of like Trump and Zelensky just did. Not a technical reason. Sometimes in business the deal falls through I was on the receiving end and didn’t appreciate it.

jml78 2y ago

I have used all the platforms personally and professionally. GCP, AWS, Azure, Oracle Cloud. I will just say that Azure seems to want to do shit different for the sake of being different. It is really annoying to write infrastructure as code for aws/gcp then go to do the same for azure and realize how dumb some of their stuff is. Just my personal experience.

gdiamos 2y ago
That’s how clouds try to lock you in, by making you use a custom tool that is different for the sake of being different. If you use standard tools you don’t have this problem. Containers running on VMs is standard. A mesh of microservices that depend on cloud queues and managed services is not. One argument against standard containers is saving dev time. You can still save dev time by using standard open source software. How many different ways are there to implement a queue or a load balancer? If you really need access to some proprietary technology then by all means use the cloud that offers it. Eg if your customer demands GPT4.5, then go with Azure. But if you need something standard, don’t get caught in the trap.
- jml78 2y ago
  I am an older guy that was building kubernetes clusters before eks, aks, gke. So I used terraform to build shit out to make it happen. Azure was 5x the code just to be different. You can try to blame terraform but if you used MS custom tooling it was no different.
- rconti 2y ago
  What about the way Terraform is a 3rd class citizen on Azure? And there are multiple ever-changing ways of doing everything, major parameters aren't supported, etc. It just makes it more difficult to deal with. Also, Azure APIs are incredibly slow.
  - re-thc 2y ago
    It’s Bicep
    - rconti 2y ago
      I'd rather avoid the trap. I use the 3 major CSPs, so I would prefer to use cross-platform tooling. Per the parent: >That’s how clouds try to lock you in, by making you use a custom tool that is different for the sake of being different. > If you use standard tools you don’t have this problem.
jiggawatts 2y ago
Conversely, doing things "the same way" as AWS would mean copying their first-generation public cloud design flaws. The overall UX of AWS is absolutely crazy. It's easy to "lose" a resource... in there... somewhere... in one of the many portals, in some region, costing you money! Meanwhile, Azure shows you a single pane of glass across all resource types in all regions. It's also fairly trivial to query across all subscriptions (equivalent to AWS accounts). Similarly, AWS insists on peppering their UI with random-looking internal identifiers. These are meaningless and not sortable in any useful way. Azure in comparison allows users to specify grouping by "english" resource group names and then resources within them also have user-specified names. The only random identifiers are the Subscription GUIDs, but even those have user-assignable display name aliases. The unified Portal and scripting experience of Azure Resource Manager is a true "second generation" public cloud, and is much closer to the experience of using Kubernetes, which is also a "second gen" system developed out of Borg. E.g.: In K8s you also get a single-pane-of glass, human-named namespaces (=resource groups), human-named workloads, etc...
- smackeyacky 2y ago
  A single pane of glass that shows all your resources that are currently choking due to hidden limitations is no flex over AWS. It is my hope I never have to use Azure ever again professionally or otherwise
iamtheworstdev 2y ago
this killed me. i dont think AWS does everything well... but Azure went hard on being different just to be different.. like region names...
stego-tech 2y ago
> I will just say that Azure seems to want to do shit different for the sake of being different. That's Microsoft's MO in a nutshell in my experience, and I say this as a recent(~5yrs ago) convert to Linux who built a career on Windows endpoints, servers, ADDS, Exchange, SCCM, you name it. It's how they achieve lock-in to their ecosystem, and it's incredibly frustrating to see how they've just layered that method of operation over and over again, decade after decade, rather than fix anything.
- Henchman21 2y ago
  Fixing things is hard. Papering thing over with free Azure credits and marketing? This is the way.

BoredPositron 2y ago

The UX not being designed by business majors.

inetknght 2y ago
You won't find that in major cloud providers.

zeroxfe 2y ago

> What do you really need out of a cloud? I need a cloud to be reliable and secure. I've used Azure extensively and it's neither. I'll take GCP or AWS over Azure any day.

foxylad 2y ago
This. We've used GCP Appengine for years and it is rock solid. Their SRE game is top level, and when there is an outage, they do a serious investigation and make it fully public, even if they screwed up badly. Including the vital "this is how we're going to stop this ever happening again". The last outage (that we noticed) was several years ago.

avereveard 2y ago

Azure tend to bite back when they upgrade their backend and everything breaks Happened twice with my kubernetes deportment, first something with node groups made them incompatible and had to recreate the cluster from scratch, then one of their scripts to rotate key access to volumes (that one has to run manually, go figure) stopped working and caused my volumes to detach from pods, and has to recreate the cluster again, and I just give up. I was super happy as well, the first two years. By years six I was fully migrated out.

FridgeSeal 2y ago

I need it to be sane, to work and to be reasonably well documented. Azure ostensibly fails outright on points 1 and 3, and limps by on 2. The products are confusing mess, there’s way too many ways to auth things, docs are garbage, tried multiple times to manage stuff via Terraform which broke far too much to be excusable, to say nothing of the dumpster fire that is their UX. I’m sure some people have either beaten it into submission, or have stockholmed themselves into putting it up with it.