Operating Intello at scale

Guidance for organizations monitoring dozens to hundreds of vendor APIs without drowning in noise or administrative overhead.

Naming and ownership

Encode vendor + surface + version in dependency names (Stripe Billing 2024-11, not API 7).
Assign domain mappings early so teams filter to their scope.
Document in your CMDB which Intello dependency (name + link from the app) backs each production integration.

Library vs BYO split

Prefer provider subscriptions for commodity SaaS APIs Intello already tracks.
Reserve BYO dependencies for gateways, forks, or specs with proprietary extensions.

Sync cadence policy

Critical paths: minimum allowed interval per plan.
Low-risk read-only specs: longer intervals to reduce worker load.
Avoid setting every dependency to the minimum interval “just because”—you’ll hit plan ceilings and dilute urgency.

Alert architecture

Default deny for PagerDuty: route only breaking + critical domains.
Use channel allow-lists on preferences to keep Slack high-signal.
Keep notify_on_breaking enabled for ops-facing rules even when min_severity is relaxed for email.

Optional automation

Most orgs live entirely in the web app. If you build internal dashboards on top of Intello, use the HTTP API sparingly: cache lists, respect rate limits, and paginate large exports—see Core resources.

Governance

Restrict admin roles to platform leads; keep integrators as members.
Rotate webhook URLs and PagerDuty keys on personnel changes.
Quarterly audit disabled dependencies—delete cruft to simplify dashboards.