News for hackers. Stuff that matters.
Re:Microservices Considered Harmful (2019) by z3t4 on Sunday June 07, 2020 @03:22PM (#23449952)
Take away the word micro from micro services. Its just a buzzword. Now you have just services. You can have just one service that handles email, chat, payroll, website - or you can break them up into independent services. Ask yourself: Does it make sense to have two different services to handle x and y. Just don't break something up because of some buzzword mantra. Maybe the public website is the bottleneck in your monolith, then it might be a good idea to put it on its own server and scaling strategy so that it doesn't bog down the other parts of the system.
Re:Microservices Considered Harmful (2019) by eldelshell on Sunday June 07, 2020 @04:44PM (#23450636)
Yeah, something in between a monolith and microservices, and I'm going to name it: services architecture.

Hmmm... I think I can do better: Service Oriented Architecture... Yeah I like this name. SOA.

Are you telling me I just invented something that's 30 years old? Bollocks!

Re:Microservices Considered Harmful (2019) by Bnshsysjab on Sunday June 07, 2020 @06:31PM (#23451550)
Can you articulate SOA without using buzzwords?

I assume it’s concepts like a dedicated TLS terminator, Single Sign on, centralised logging, etc?

Re:Microservices Considered Harmful (2019) by m463 on Sunday June 07, 2020 @07:48PM (#23452129)
programs communicating with remote programs.

People will try to quibble that all this is new stuff, but really it is just new names for old ideas.

https://en.wikipedia.org/wiki/Remote_procedure_call

Re:Microservices Considered Harmful (2019) by ismayilzadan on Sunday June 07, 2020 @04:04PM (#23450310)
Agreed. I was going to add my own separate comment but noticed yours. I think the main issue is, most companies nowadays end up creating "nano" services. A normal "service" should be built and then if functionality/team gets too big, then you break it down into another 2-3 smaller services. Again, service code already should have been written in a clean, modular code style so it can be broken down easily too without rearchitecting whole stack.
Re:Microservices Considered Harmful (2019) by ShellfishMeme on Sunday June 07, 2020 @05:58PM (#23451247)
But is it really true that so many companies create too small services? I admit I haven’t seen too many architectures from the inside but from my experience the problem is usually the opposite: adding more and more things to a monolith that already has serious complexity and scalability issues because it’s the convenient place and is faster to develop. Most of the times I heard about the nano services problem was from articles warning of using services and encouraging monoliths instead.
Re:Microservices Considered Harmful (2019) by ismayilzadan on Monday June 08, 2020 @07:27AM (#23455613)
It is true for companies that recently trying to adopt this "microservices" catchphrase. Just because it is a trend and somebody sold kubernetes/cf/some-other-cloud-solution to the company, they end up overusing it and creating a new microservice for something that could just be a separate package in already existing microservice. I am definitely not encouraging monolith architecture, they are horrible to work with, but I am also not suggesting 1500+ microservices for a company like Monzo (just considering their lack of functionalities compared to other new banks, 1500+ microservices sound like nanoservices instead to me)
Re:Microservices Considered Harmful (2019) by jayd16 on Sunday June 07, 2020 @11:41PM (#23453375)
Yeah, I think this is an important point. Micro is supposed to imply that you should feel free (not obligated) to make small services.

People ended up treating it like every library project should be separated by a network call.