multiple microservice deployments from a monorepo

[lantz moore]

i am trying to build and deploy two separate microservices (api and client) from a monorepo.

i added four environments: api-qa, client-qa, api-prod, and client-prod.

definitions:
  steps:
   - step: &build-api
       name: build-api
       condition:
         changesets:
           includePaths:
             - "shared/**"
             - "api/**"
       script:
         - echo build api
   - step: &build-client
       name: build-client
       condition:
         changesets:
           includePaths:
             - "shared/**"
             - "api/src/entities/**"
             - "client/**"
      script:
        - echo build client
    
    - step: &deploy-api-qa
        name: deploy-api-qa
        deployment: api-qa
        condition:
          changesets:
            includePaths:
              - "shared/**"
              - "api/**"
        script:
          - echo deploy api to qa
     - step: &deploy-client-qa
         name: deploy-client-qa
         deployment: client-qa
         condition:
           changesets:
             includePaths:
               - "shared/**"
               - "api/src/entities/**"
               - "api/**"
         script:
           - echo deploy client to qa

     - step: &deploy-api-prod
         name: deploy-api-prod
         deployment: api-prod
         trigger: manual
         condition:
           changesets:
             includePaths:
               - "shared/**"
               - "api/**"
          script:
            - echo deploy api to prod
     - step: &deploy-client-prod
         name: deploy-client-prod
         deployment: client-prod
         trigger: manual
         condition:
           changesets:
             includePaths:
               - "shared/**"
               - "api/src/entities/**"
               - "api/**"
         script:
           - echo deploy client to prod

pipelines:
  default:
   - step: *build-api
   - step: *build-client
   - step: *deploy-api-qa
   - step: *deploy-client-qa
   - parallel:
     - step: *deploy-api-prod
     - step: *deploy-client-prod

this *almost* works.  i do end up with four different environments, however, the "promote" button acts weirdly.  there's no promote button on api-qa and the promote button on client-qa wants to promote to api-prod.

what am i doing wrong here?  thanks in advance.

[Florian PEREZ · Accepted solution]

Hello,

Since this question is very technical I suggest you to post this on the developper part of the forum:

https://community.developer.atlassian.com/

I think you'll get more answers and from a better quality!

Hope this help

[Ilya Yablochnikov]

The problem is that the steps are executed in order. Since `api-prod` comes after `deploy-client-qa`, when it comes time to promote from "qa" to "prod", it executes the first step in "prod".

I sort of fixed this by making putting all "qa" steps under one parallel step and all "prod" steps under another parallel step.

The problem is that I have to finish deploying all the "qa" steps before I can start promoting to "production". But at least I can choose which "qa" step to run, without having to run them in the exact sequence they appear in bitbucket-pipelines.yml.

Unfortunately there does not seem to be a way to be choose to deploy to one "qa" deployment and then promote it to any one of the "prod" deployments.

So in my case, I gave up and declared all my environments as "qa" and just have to be smart about which order to deploy things. All I wanted was an easy way to push my artifact to different servers, that's why I even bothered with Bitbucket Pipelines.

[koti r]

Hi, Iantz moore

I have one query related to this. What I have to do if I want to re-use the deployment environments like TEST, STAGING & PROD, to deploy 20+lambdas. Should I have to define all these lambdas in definitions section of the pipeline?. Please suggest the best way to deploy into multiple environments in multiple times.