[Erin McNulty (erin2722)]

Default the chart and GKE examples to two dedicated Brainstore automation writer replicas. Existing templates isolate the automation loop from regular writers and publish the dedicated Service endpoint to the APIs. Update the values guidance and add coverage for the default replica count, loop isolation, and API routing.

This adds two pods at the existing resource requests (32 CPU and 128Gi memory each). Explicit replica overrides and retained values remain authoritative. No chart version or image tag changes are included.

Input matrix:

• Omitted override with current chart defaults: two replicas, dedicated Service and API endpoint, regular writers exclude automations.
• Explicit 0, null, empty replica count, and missing pool map in reused values: preserve existing disabled/fallback behavior. Partial reused maps keep existing default merging.
• Supported release tags, SHAs (including numeric short SHAs), prereleases, and custom tags: existing validation and rendering behavior remain unchanged. Release tags older than v2.16.0 require an explicit zero pool; legacy-image CI and startup-gate tests now set that override.
• AWS, Google, Azure, and minimal cloud configuration: render with the enabled pool. Existing Azure storage requirements still apply.
• Enabling/disabling the pool retains the existing loop-selector and routing behavior; removing the pool still requires API routing coordination. Upgrades retaining an explicit 0 do not automatically enable the pool.

Validation: ./test.sh (424 unit tests, Python startup-gate tests, all CI cloud renders, strict Helm lint); separate enabled-pool renders for AWS, Google, Azure, and minimal configuration.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-30T16:12:27.882308Z	61e1a91	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review" or "Codex (@codex) security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 61e1a916c1

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "Codex (@codex) address that feedback".

[chatgpt-codex-connector]

Keep automation writers opt-in until rollout is coordinated

When an install or upgrade adopts this new default, the chart immediately publishes BRAINSTORE_AUTOMATION_WRITER_URL and configures regular writers to exclude automations, but the API startup gate does not wait for the automation-writer Deployment. As documented in braintrust/README.md, enabling the pool during an upgrade can temporarily run automations on both pools, while an unschedulable new pool—each replica requests 32 CPU and 128Gi—can instead leave no Ready process handling automations after the regular-writer rollout. Keep the default at zero or add sequencing/readiness protection before enabling it automatically.

AGENTS.md reference: AGENTS.md:L99-L101

Useful? React with 👍 / 👎.