WIP: Playbook self-managed operator docs

[tylerhutcherson]

MOD-19679 — Self-managed Redis Agent Playbooks operator documentation.

Adds setup and administration docs under /operate/iris/agent-playbook/. An operator can prepare Redis and provider dependencies, install the published Helm chart, provision a playbook and scoped application keys, and hand a working connection to a developer.

Scope

• Ten pages: operator landing, self-managed overview, deployment planning, prerequisites, configuration, deployment, authentication/access, API examples, operations, and the generated public control-plane API reference.
• Chart 0.1.0-alpha.3 and its packaged images are the verified baseline. Configuration covers separate metadata/content Redis, provider Secrets, Helm values, and private CA trust.
• Operations cover health/troubleshooting, key rotation/revocation, suspension/deletion, Secret checksum rollouts, backup responsibilities, and chart upgrades.
• Follows the Agent Memory operator-docs structure, with Iris navigation and developer handoffs for setup, access, and retrieval settings. Content authoring and agent/framework integration stay in the parent developer PR.
• The follow-up removes 115 words across five pages, makes credential loading an explicit setup step, and keeps the API walkthrough focused on provisioning, publisher/reader keys, and health. Publisher connections hand off to the developer quickstart; reader connections hand off to agent integration. Detailed service configuration stays in Configuration. Operator procedures use standard Bash/YAML fences and link to the parent's canonical TCE programs; no separate SDK example or rendering system is introduced.

This PR is self-managed only. Cloud onboarding will follow when available. Extraction deployment, service changes, and SDK/package upgrades are outside this PR. Backup responsibilities are documented; a qualified restore procedure remains open.

Validation

• Fresh onboarding smoke test passes against the existing licensed Helm deployment: displayed bootstrap/create/read/key/health commands, the canonical TCE publisher quickstart, reader retrieval with denied authoring, and the canonical TCE reader MCP client. The disposable playbook and keys are cleaned up. No additional service instance is started.

• Installed the published chart with Helm 3.22.0, an entitled Iris suite license, separate persistent Redis deployments with verified TLS, and real OpenAI embeddings. All three service deployments became ready.

• All four chart tests and ten application smoke checks pass, including provisioning, FAQ/skill publication and retrieval, tenant/action boundaries, rotation/revocation, suspension/resumption, and deletion.

• Same-version Helm reapply preserves the four generated service credentials. Storage Secret/checksum updates roll all three services while preserving the playbook, content, and reader access. Port-forward reconnect guidance is tested after pod replacement.

• All 33 remaining Bash and six YAML snippets parse; Linux/macOS checksum commands and CA Secret setup were exercised. The generated reference matches the pinned release contract. All retained executable snippets and frontmatter are unchanged; the settings example is replaced by a Configuration link.

• Current with the updated developer branch (d7293c339), including its upstream Hugo build fix. The standard Hugo build passes without a temporary compatibility mount. All 21 pages / 136 internal links and static anchors across the stack resolve, with clean Markdown exports. Developer validation passes 126 scoped checks, strict TypeScript compilation, source-lock validation, dependency checks, and workflow lint; desktop/mobile TCE copies and API model anchors pass. Fresh Chrome checks also cover all ten operator pages at desktop/mobile widths, including the control-plane reference model, introduction, and OpenAPI download. The five reference/source checks pass after cleanup.

The Kubernetes smoke test runs AMD64 release images through Docker Desktop emulation on an ARM64 host. It verifies this installation, not a production platform/support matrix or native ARM qualification. Existing warnings remain in unrelated site sections. Incremental Hugo previews can require restarting for complete-program copy templates; clean preview builds pass.

Review readiness and remaining work

The core self-managed operator documentation is ready for human review now. These remaining items are planned for the next one to two weeks, with release-dependent additions handled in follow-up PRs:

• Approve launch chart/images, update embedding connection configuration, and regenerate the public control-plane reference together.
• Confirm clearing and extraction availability, then document supported source connections and worker/queue setup.
• Confirm and document identity-provider and browser-administration setup for the launch release.
• Qualify and document supported Kubernetes/platform versions, production sizing, and HA guidance.
• Qualify ingress/TLS, network controls, and minimum Redis ACL privileges.
• Verify cross-version upgrade compatibility and publish a tested backup/restore procedure.
• Resolve or document the shared Hugo incremental-preview issue. Restarting the preview is the current workaround; clean builds pass.

Stack and publication

• Parent: #4249 — WIP: Playbook developer docs, targeting main.
• This PR: MOD-19679-playbook-operator-docs → MOD-19391-playbook-developer-docs. The child diff contains only operator docs, navigation/cross-links, and operator review notes.
• Merge the developer PR first, then retarget/rebase this PR onto main and verify the remaining diff.

Both PRs stay draft/WIP. No publication gate is added. Merging into main publishes the docs; draft state does not hide merged pages. Keep both unmerged until launch publication is approved.

[github-actions]

🧠 Redis Memory

Found 5 related items from repository history:

• PR — WIP: Playbook developer docs

  MOD-19391 — Publish Redis Agent Playbook documentation on redis.io.

• PR — WIP: Playbook developer docs

  Adds the Agent Playbooks developer docs under /develop/ai/context-engine/agent-playbook/ for docs-team and product review. The guides cove

• PR — MOD-18528: Document Playbook bindings and RAM service-subject authentication

  Summary

• PR — MOD-17747: Add self-managed LangCache documentation

  Summary

• PR — MOD-18825: Correct Iris self-managed docs for RAM 0.7.0 and LangCache 0.0.1

  Ticket: MOD-18825 · Control Plane OpenAPI text: [redislabsdev/iris#1648](https://github.

Memory updated at 14dcc3e

[github-actions]

Staging links:
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/develop/ai/context-engine/agent-playbook/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/develop/ai/context-engine/agent-playbook/concepts/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/develop/ai/context-engine/agent-playbook/quickstart/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/develop/ai/context-engine/agent-playbook/retrieval/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/api-examples/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/authentication/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/configuration/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/control-plane-api-reference/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/deploy/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/operations/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/plan-deployment/
https://redis.io/docs/staging/MOD-19679-playbook-operator-docs/operate/iris/agent-playbook/self-managed/prerequisites/

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[mich-elle-luna]

Hi Tyler, this is looking good so far. But, I've merged in lots of changes since you branched, so here is the info to get this branch to appear with the new look/feel and nav menu from claude:

"Nothing in their pages needs to change. Their branches were cut on Sept 30, before the Iris work merged, and staging builds each branch as-is. Both branches are 133 commits behind main, so they don't have the new nav or dark theme.

What they need to do: Merge main into both branches (or rebase onto it) and push. Their new pages are under develop/ai/context-engine/agent-playbook/ and operate/iris/agent-playbook/. Everything under those two folders gets the Iris sidebar and dark theme automatically, with no frontmatter or extra setup.

What they'll run into:

Developer branch (MOD-19391-playbook-developer-docs): A dry-run merge with main is clean.
Operator branch (MOD-19679-playbook-operator-docs): One conflict, in content/operate/iris/_index.md. They added an Agent Playbooks tile and reworded the intro. Meanwhile main added a console screenshot, switched the tiles from color swatches to icons, and added lists of links. Resolving it means keeping main's version and re-adding their tile and wording.
Their new tile uses a color swatch (color="bg-blue-300"), while the other tiles now use icons. No playbook icon exists yet. They could use icon-redis-iris-64-duotone.png for now, or ask for a new icon."

thank you!