Azure Service Plans and Non-Productive Slots

Question

I'm looking for best practice when it comes to azure service plans in microservice architecture. We have a series of microservices, each of which is completely independent of each other both in terms of capacity, resources, developers, and the overall architecture. It goes without saying that if one service is experiencing problems, others should not be affected if they do not interact with the problem service. These services are hosted on Azure.

My question is about Service Plans and how they should relate to the development / installation environment. So far, we have created a service plan for our microservice, name it PersonService. Thus, we will create a PersonService service plan, and then the default slot will be production (personal service), and then we will have another intermediate segment (personal service) to meet the needs for production / testing. All of them will be serviced according to the same tariff plan.

Today I got the scary idea that if a developer deploys some terrible error in a process that consumes the entire processor and / or memory, then the production slot will starve from these resources, and, in fact, the intermediate environment will affect the response time production.

May I think it will be so? How do you guys recommend installing this to avoid this problem? Thanks

Answer 1

Yes, you are right, if the personnel management process begins to consume a significant part of the server’s basic resources, this will affect the user’s service.

Avoiding this very much depends on your current setup and your priorities. Adding a dev / test / staging service plan is the easiest way. This leaves the production plan exclusively for the finished production code. With deployment slots there just to make it easy to switch between versions (and quick rollback if you understand that something in the production is broken)

, , . , Azure , , . , .

. ", " CI - 30 ! - . , Azure Automation, - ( , ! - /)