fix: look up and pass stage 1 name

[samrose]

1. After stage 1 completes, a new "Find stage 1 AMI" step queries AWS for the AMI by:
   • postgresVersion tag = ${version}-stage1
   • sourceSha tag = ${GIT_SHA} (unique to this build)
2. Exports STAGE1_AMI_ID to the environment for the next step
3. Stage 2 now receives -var "source_ami=${STAGE1_AMI_ID}"

Summary by CodeRabbit

• Chores
  • Pipeline now automatically discovers and validates the appropriate base image for stage 2 builds, reducing manual configuration and improving build reliability.
  • Postgres platform mappings updated to newer patch releases for 17 and 15 series (including the OrioleDB variant), ensuring deployments use up-to-date minor/patch versions.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

📝 Walkthrough

Walkthrough

The PR adds a step to the AMI release GitHub Actions workflow that discovers a stage 1 AMI ID by querying AWS for an image tagged with postgresVersion and sourceSha in the region derived from the Packer config, validates it, and exposes it as STAGE1_AMI_ID. The stage 2 Packer build step is updated to pass that AMI ID as the source_ami variable. Separately, ansible/vars.yml bumps three Postgres release strings (postgresorioledb-17, postgres17, postgres15) to new patch versions.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Possibly related PRs

• feat: add branch-based versioning for PR AMI builds #1910: Touches the same ami-release-nix.yml workflow and the two-stage AMI build flow.
• fix: somehow dropped needed suffix #2026: Updates the same ansible/vars.yml mapping key (postgresorioledb-17) with an OrioleDB version suffix.
• chore: bump version to correct new version for potential release #1991: Bumps entries in ansible/vars.yml for Postgres package versions.

Suggested labels

ci, infrastructure, ami

Suggested reviewers

• hunleyd
• LGUG2Z

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Title check	❓ Inconclusive	The title 'fix: look up and pass stage 1 name' is vague and doesn't clearly convey what is being looked up or passed. The phrase 'stage 1 name' is ambiguous—it could refer to an AMI ID, a name, or other identifiers.	Revise the title to be more specific, such as 'fix: look up and pass stage 1 AMI ID to stage 2 build' to clearly indicate the primary change.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description check	✅ Passed	The description clearly outlines the three key changes: querying AWS for stage 1 AMI using specific tags, exporting STAGE1_AMI_ID, and passing it to stage 2. It provides sufficient detail about the implementation.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch sam/source-ami-fix

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Fix all issues with AI agents

In @.github/workflows/ami-release-nix.yml:
- Around line 97-99: The REGION extraction using grep '^region=' is brittle and
can yield an empty value; update the REGION assignment (where PG_VERSION and
REGION are set) to use a more flexible pattern that accepts optional spaces and
quotes (e.g., match lines like region = "us-east-1") and trim surrounding
quotes/whitespace, then add a guard after REGION is computed to validate it is
non-empty and exit with a clear error message if missing so the workflow fails
fast with a helpful diagnostic.
- Around line 102-110: The describe-images query that sets STAGE1_AMI_ID
currently uses Images[0] which is unordered; update the AWS CLI --query to sort
images by CreationDate and select the newest one deterministically (e.g. use
sort_by(Images,&CreationDate)[-1].ImageId or equivalent) so STAGE1_AMI_ID always
picks the most recently created AMI matching the filters; modify the command
that sets STAGE1_AMI_ID accordingly.