The connectors in the box — Azure, Microsoft 365, AWS and GCP (the last two in Beta), Kubernetes, MuleSoft and Docker, with a set of SaaS platforms still marked coming soon — cover the bulk of what most teams run. But the underlying CloudMonitor architecture — scanner pulls metadata, normalizes into a common resource & finding model, runs rules, renders in the dashboard — doesn't care about the source. If a system has an authenticated API, we can build a connector for it. Custom connectors are built to order; none has shipped to a customer yet, so read this as how we would do it rather than a catalog.
What "we can connect to it" means concretely
- Any REST or GraphQL API with OAuth2, API-key, bearer-token, or mTLS auth.
- Any SaaS platform that exposes admin / observability endpoints — HRIS, finance, identity, observability, ticketing, monitoring-of-monitoring.
- Any internal API that our scanner can reach over the internet — we can connect to your services and pull the metrics that matter to your operations team, not just generic ones.
What it does not mean: there is no on-premises agent and no private-network link today, so a system that is only reachable from inside your network is out of reach for now.
The pattern
- You give us API docs (or point us at OpenAPI / Swagger) and a sample read-only credential.
- We map the resource model into our normalized schema (the same Resource + Finding model the Azure / AWS / GCP connectors use).
- We write a small scanner module that runs on the same schedule as the built-in ones.
- It is enabled for your tenant first; whether it is ever offered to other customers is your call.
Examples of what a custom connector could surface
- An internal "feature flags as a service" platform — we'd surface stale flags, flags rolled to 100% for >90 days, and flags with no owner.
- A CI vendor not in the box (Buildkite, CircleCI, Jenkins) — we'd surface broken pipelines, flaky tests by frequency, queue times above your own target.
- A homegrown deployment system — we'd surface failed deploys, rollbacks, drift between declared and running versions.
- A network appliance with an admin API — we'd surface config drift, firmware lag, certificate expiry.
What we won't do
- We won't build a connector that needs write access. Nothing in CloudMonitor writes to a connected system today.
- We won't build a connector that reads end-user personal data unless you have a clear, lawful purpose and we've agreed it in writing first.
- We won't resell a custom connector built for you without your written sign-off.
If you have a system in mind and want to know if it fits, the fastest path is to get in touch and send the API docs.