Somehow, when a new environment gets deployed from a PR or branch the last step fails with the message:

Not making progress trying to resolve environmental context. Giving up.
[Error at /staging-company-stack/Stack] Found zones: [] for dns:<http://dev.example.com|dev.example.com>, privateZone:undefined, vpcId:undefined, but wanted exactly 1 zone
  Annotations.addMessage (/tmp/seed/source/node_modules/aws-cdk-lib/core/lib/annotations.ts:64:25)
  Annotations.addError (/tmp/seed/source/node_modules/aws-cdk-lib/core/lib/annotations.ts:36:10)
  Function.getValue (/tmp/seed/source/node_modules/aws-cdk-lib/core/lib/context-provider.ts:83:31)
  Function.fromLookup (/tmp/seed/source/node_modules/aws-cdk-lib/aws-route53/lib/hosted-zone.ts:76:65)
  ReactStaticSite.lookupHostedZone (/tmp/seed/source/node_modules/@serverless-stack/resources/src/StaticSite.ts:511:39)
  new StaticSite (/tmp/seed/source/node_modules/@serverless-stack/resources/src/StaticSite.ts:101:28)
  new ReactStaticSite (/tmp/seed/source/node_modules/@serverless-stack/resources/src/ReactStaticSite.ts:43:5)
  new PlatformStack (/tmp/seed/source/stacks/PlatformStack.ts:11:18)
  Object.main (/tmp/seed/source/stacks/index.ts:18:3)
  Object.<anonymous> (/tmp/seed/source/.build/run.js:94:16)
  Module._compile (internal/modules/cjs/loader.js:1063:30)
  Object.Module._extensions..js (internal/modules/cjs/loader.js:1092:10)
  Module.load (internal/modules/cjs/loader.js:928:32)
  Function.Module._load (internal/modules/cjs/loader.js:769:14)
  Function.executeUserEntryPoint [as runMain] (internal/modules/run_main.js:72:12)
  internal/main/run_main_module.js:17:47
Found errors
Error: Found errors
    at StackCollection.processMetadataMessages (/tmp/seed/source/node_modules/aws-cdk/lib/api/cxapp/cloud-assembly.ts:270:13)
    at CdkToolkit.validateStacks (/tmp/seed/source/node_modules/aws-cdk/lib/cdk-toolkit.ts:549:12)
    at CdkToolkit.selectStacksForDiff (/tmp/seed/source/node_modules/aws-cdk/lib/cdk-toolkit.ts:528:10)
    at processTicksAndRejections (internal/process/task_queues.js:93:5)
    at CdkToolkit.synth (/tmp/seed/source/node_modules/aws-cdk/lib/cdk-toolkit.ts:412:20)
    at initCommandLine (/tmp/seed/source/node_modules/aws-cdk/bin/cdk.ts:268:9)
It seems you are configuring custom domains for you URL. And SST is not able to find the hosted zone "<http://dev.example.com|dev.example.com>" in your AWS Route 53 account. Please double check and make sure the zone exists, or pass in a different zone.
ERROR: Error: Found errors
ERROR: Failed to generate artifact for downstream stage staging.

This is our seed configuration. Development pipeline • Github

dev

branch ->

dev

environment • on PR activity -> new environment • on branch activity -> new environment Staging pipeline • Github

main

branch ->

staging

environment Production pipeline • Manually promote

prod

environment It seems that seed picks up the wrong AWS credentials during the last step. 📦 Build:

$ npx sst build --stage pr9 --verbose
...
Resolving default credentials
Retrieved account ID 24********** from disk cache
Setting "CDK_DEFAULT_ACCOUNT" environment variable to 24**********
context: {
  'hosted-zone:account=24**********:domainName=<http://dev.example.com:region=us-east-1|dev.example.com:region=us-east-1>': { Id: '/hostedzone/Z02179983PO**********', Name: 'dev.example.com.' },
...

🗑️ Cleanup:

$ npx sst build --stage staging --verbose
...
Resolving default credentials
Looking up default account ID from STS
Default account ID: 01**********
Setting "CDK_DEFAULT_ACCOUNT" environment variable to 01**********
context: {
  'hosted-zone:account=24**********:domainName=<http://dev.example.com:region=us-east-1|dev.example.com:region=us-east-1>': { Id: '/hostedzone/Z02179983PO**********', Name: 'dev.example.com.' },
...

On the settings page I can see that: • Dev AWS IAM = custom (24**********) • Prod AWS IAM = custom (its own) • Staging AWS IAM = use default • The PR AWS IAM = custom (24**********) • The Branch AWS IAM = custom (24**********) • Default settings = 01********** I would assume that the last step is somehow using the default settings for PR and Branches.

Hey @Daniel Gato, the last step should use Staging AWS IAM. What it’s trying to do is running

sst build

for the downstream stage.

It’s using the default IAM. Staging uses the default as well. So indirectly, it is using staging’s IAM.

Yeah I see what you mean. Just curious, would it work if you relied on the stage name (ie.

app.stage

) instead of

SEED_PULL_REQUEST_NUMBER

to determine whether you are building for staging or prod?

the quickfix we did right now is:

const isPullRequestOrBranch =
    (isPullRequestDeploy || isBranchDeploy) && !isStaging;