Your runbook is only a control if the run has to read it.
Confluence is a first-class knowledge connection provider in NopsAI, alongside Notion and wiki sources. Pages are converted from Confluence storage format into structured content a run can use, which is what turns documentation into something a pipeline can be held to.
Confluence inside a governed run.
How Confluence participates once it is connected — and what the platform does with what it returns.
Native storage-format conversion
Confluence pages are converted from their storage format into prompt-ready content, preserving structure and extracting referenced assets rather than flattening a page into unusable markup.
Referenced, not copied
A pipeline attaches knowledge context by reference. The page stays where its owners maintain it, and the run picks up what the reference resolves to instead of a copy that silently ages.
Blocking when it needs to be
Marking an entry required turns reference material into an enforcement point: execution is validated against that context before it proceeds, including for plain script steps.
Scoped to the run that needs it
Knowledge context can be attached at pipeline, step or task level, so a production policy applies to the deployment step without being fed into every unrelated task.
The controls that make it safe to leave connected.
An integration is easy to add and hard to bound. These are the parts that decide whether it stays reviewable six months later.
Access uses a stored credential reference
The connection authenticates with a credential from the encrypted registry. Configuration carries the reference, and access by the connection is logged like any other credential consumer.
What was read is recorded
The context resolved for a run is part of its evidence, so a later review can establish which version of a policy the run was actually held to.
Content sharing rules still apply
Knowledge content reaching a model is governed by the same content-sharing settings as workspace content, so connecting a space does not implicitly widen what may be sent to a provider.
Unavailable context fails closed
Under the default governance level, required context that cannot be resolved stops the run. A policy that silently failed to load would be worse than no policy.
Owned by Git, reviewed like code.
Configuration is a reviewable file rather than a form someone filled in once. Credentials appear as references; the values live in the encrypted registry.
name: production-deploy
container_image: bitnami/kubectl:latest
governance_level: strict
knowledge_context:
- kind: policy
ref: platform/production-change-policy
required: true
- kind: runbook
ref: platform/deploy-runbook
steps:
- name: plan
goal: Summarize the change and check it against the policy.
- name: approve
depends_on: [plan]
approval:
teams:
- platform/prod
allow_self_approval: falseWorkflows where Confluence shows up.
The same integration, seen from the workflow rather than from the connection.
A customer-reported bug needs to ship as a hotfix
The fix is small, but no single durable trail connects the complaint, the code change, the test result and the emergency approval.
Production incident at 2 a.m.
At 2 a.m. the hard part is finding what changed and deciding whether rollback is safe — fast, without losing the evidence trail.
Preparing a release
Scope mismatch, unmerged PRs, failed gates and an outdated runbook can all cause a late surprise.
What teams ask about Confluence.
Does connecting Confluence send our whole space to a model?
No. A run resolves the specific entries its pipeline references, and what may then be sent to a provider is governed by the pipeline's content sharing settings.
What does marking context required change?
It makes validation blocking for whatever scope it is attached to. Even a direct script step is validated against that context before it runs, and under the default governance level a non-allow or unavailable review stops the run.
Can Confluence pages be used with other sources?
Yes. Confluence, Notion and wiki connections can all supply knowledge context, and a pipeline can attach entries from more than one alongside context stored in the configuration repository.
Map your Confluence workflow end to end.
Bring one real workflow. We will map its trigger, tools, approvals, runtime boundary and evidence requirements with you.