Skip to content
Patch What Matters

CONNECTORS, read 10 Nov 2026

Plan connectors from your own list of sources by kind, because vendor counts of connectors are not comparable with each other

All guidesGlossary

Why the headline count misleads

The Connector Census on this site records, for each profiled platform, the connector count the vendor documents on its own pages, with the page. On 2026-11-10 the figures were: Rapid7 500, ArmorCode 350, Phoenix Security 320, Tenable 300, Brinqa 260, Seemplicity 190, Nucleus 183, Zscaler 150, Hackuity 130, Qualys 100 and Vulcan Cyber 100. ServiceNow's Vulnerability Response and Cisco's Vulnerability Management documentation gave no single count.

These numbers are not comparable one to one. One vendor counts each tool; another counts each integration direction; a third counts each product module. A catalog of 500 that lists a dozen of the tools you use is less useful than a catalog of 130 that lists all of them. Nucleus's own home page claims 200+ connectors where its integrations page lists 183, which is a small example of how counts drift between pages.

Start from your sources

List every system that produces findings or context for the program, in the seven kinds the census uses: infrastructure scanners, application security, cloud and container, pentest and bug bounty, asset and CMDB, ticketing and workflow, and threat intelligence. For each, write the product name and version, who administers it, how it exports (API, file, database) and how often. Mark each as day one, month three or later.

Count the day-one list. Ten instances of one scanner are one connector type with ten configurations, so count types and instances separately, because instances drive setup time and types drive the connector fit.

Check by name, kind by kind

For each day-one source, check the vendor's integration page or documentation for the product by name, and record the page. The census lists the examples each vendor names per kind. For infrastructure scanners, Brinqa names Tenable, Qualys, Rapid7, Tanium, CrowdStrike and Microsoft Defender for Endpoint; Hackuity names Qualys VMDR, Tenable VM, Nessus, Rapid7 InsightVM, Greenbone and Outpost24. Those are the vendors' statements, taken from their pages.

Pay attention to the kinds where coverage is thinnest. In the census, the number of pentest and bug bounty sources named per vendor is small: Qualys, Rapid7 and ServiceNow name none on the pages read, while Brinqa names five and Seemplicity six. If your findings come from penetration test reports, check that kind first.

Treat the vendor's examples as a starting point, not a complete list. The census records only the products a vendor names on the pages read, so a missing name is not proof of no support. Ask for the full list in writing and date the answer.

Universal and custom adapters

A documented adapter for unlisted sources matters for the long tail: an in-house scanner, a vendor with a CSV export, a niche tool. The census records whether each vendor documents one. On 2026-11-10 it found a documented universal or custom adapter for Brinqa, Cisco, Hackuity, Nucleus (FlexConnect, by Nucleus's own description), Phoenix Security, Qualys, ServiceNow, Tenable, Vulcan Cyber and Zscaler. None found on 2026-11-10 for ArmorCode, Rapid7 and Seemplicity.

Brinqa's documentation describes a connector framework covering APIs, database extracts and CSV, JSON and XML files, plus a schema-flexible CSV connector. Whatever the platform, ask how an adapter is built, who maintains it and whether it supports incremental updates, so a nightly import does not reload everything.

What a connector actually does

A connector is more than a pipe. Ask what it pulls (findings only, or assets, tags and ownership too), how it authenticates, how often it syncs, how it handles deletes and resolved findings, and what it does when the source API changes. Ask for the last three connector fixes, if the vendor publishes release notes.

For ticketing connectors, ask whether the sync is bi-directional: closing the ticket should close the finding, and a re-scan that finds the issue still open should reopen the ticket. Test that in the proof of concept, not in the demo.

Finally, plan the order of onboarding. Connect the asset source first, then the largest scanner, then ticketing, then threat intelligence feeds such as KEV and EPSS, then the long tail. Each step gets its own check: record counts against the source, a sample traced by hand, and a sign-off from the source's administrator.

Monitor connector health

A failed connector removes a source from the ranking without any visible error in the queue. Put connector last-success time, record counts and error rates on the program's own dashboard, with an alert. Count connector uptime as a coverage input in the measuring guide.

Finally, record the list of day-one connectors in the contract or the statement of work, by product name and version. A catalog count in a sales deck is not a commitment.

Sources

  1. Brinqa, connectors (vendor page), read 10 Nov 2026
  2. Brinqa documentation, connectors, read 10 Nov 2026
  3. Nucleus, integrations (vendor page), read 10 Nov 2026
  4. ArmorCode, integrations (vendor page), read 10 Nov 2026

Last reviewed 10 Nov 2026.