Skip to main content

Integrations

Integrations are how the CyberOptix CTEM Platform learns about things it cannot reach on its own. Almost every section depends on one: Cloud Exposure needs a cloud account, Application Exposure needs source control, Endpoint Security needs an EDR vendor.

An entitled section that looks empty is usually an integration that has not been connected yet.

What can be connected​

Cloud​

VendorContributes
AWSResource inventory, CloudTrail activity, CloudWatch logs, GuardDuty findings, posture assessment
AzureResource inventory, activity and resource logs, Entra ID, sign-in, provisioning, directory audit and risk detection logs, Defender, Intune, Power Platform, posture assessment
GCPResource inventory, audit logs, posture assessment

Each is read-only. Nothing is deployed into your account and nothing is changed.

The one exception to a read-only grant is Power Platform. The CyberOptix CTEM Platform only reads, but Microsoft requires an admin-level registration to read Power Platform inventory. See Power Platform.

Source control​

VendorContributes
GitHubRepositories, Dependabot alerts, code scanning alerts, secret scanning alerts
GitLabProjects and branches, SAST and SCA scanning
Azure DevOpsRepositories and pipeline activity

See Source control for what each needs and how scanning is turned on.

Endpoint​

VendorContributes
CrowdStrikeAlerts, devices, vulnerabilities, event stream
SentinelOneThreats, agents, activities, deep visibility
Carbon BlackDevices, alerts, vulnerabilities, watchlists
Microsoft DefenderDevices, vulnerabilities, software inventory, recommendations

Network and security appliances​

VendorContributes
Palo AltoTraffic, threat and URL logs; configuration and review; host and network discovery; NAT policies
FortinetTraffic, threat and event logs; configuration and review; host and network discovery; NAT policies
CiscoASA and FTD logs; ASA configuration and review; host and network discovery; NAT policies
Cisco IOS / IOS-XECatalyst switch and router configuration and review; VLANs, SVIs and subnets; connected endpoints from the MAC address table, ARP, CDP and LLDP
Cisco NX-OSNexus switch configuration and review through NX-API; VLANs, SVIs, subnets and every VRF's routes; connected endpoints from the MAC address table, ARP, CDP and LLDP
Cisco MerakiConfiguration and review; device, client, VLAN and published service discovery
JuniperSRX configuration and review; host and network discovery; NAT policies
Extreme NetworksSwitches, access points and clients through ExtremeCloud IQ and Site Engine
F5 BIG-IPConfiguration, TLS profile and certificate review; virtual servers, pools, self IPs and VLANs
InfobloxNetworks, DHCP leases, fixed addresses and DNS records
NetBoxPrefixes, VLANs, IP addresses, devices and interfaces

The appliance integrations do more than collect logs: configuration review assesses the device itself, and host and network discovery uses what the firewall already knows about your network - which often sees further than a Scanner can reach. An IPAM is the documented network, which is how Network coverage finds the subnets nothing scans.

Private cloud and containers​

VendorContributes
VMware vSphereDatacenters, clusters, ESXi hosts, virtual machines, templates, port groups
Canonical MaaSBare-metal machines, pools, zones, networks
Proxmox VEClusters, nodes, virtual machines, containers, templates, networks
NutanixClusters, AHV hosts, virtual machines, images, subnets
Microsoft Hyper-VHosts, failover clusters, virtual machines, virtual switches
Canonical LXDServers, clusters, virtual machines, system containers, images, networks
KubernetesSelf-hosted Kubernetes and OpenShift clusters: nodes, workloads, pods, containers, images, services, ingresses, routes
DockerDocker hosts and Swarms: nodes, containers, services, images, published ports
RancherEvery managed cluster: nodes, workloads, pods, images, services, ingresses
PortainerEvery managed environment: Docker hosts, Swarms and Kubernetes clusters with their containers, stacks and images

These land in Cloud Exposure beside the public cloud's resources, by resource type: a Proxmox VM sits in Compute > Instances next to an EC2 instance, a self-hosted cluster in Compute > Kubernetes next to EKS. All of them are reached through an Integrator.

Directory​

VendorContributes
Active DirectoryUsers, groups and computers over LDAP; account posture review; computers matched to hosts and endpoints

Read through an Integrator with an ordinary domain account. See Active Directory. The users, groups and computers land in Identity Exposure beside the cloud directories.

Ticketing​

VendorContributes
JiraTwo-way issue sync, reconciliation, ticket automation

Connecting one​

Integrations → Available Integrations, choose a vendor, follow its flow. What is needed differs:

Screenshot pending

SCREENSHOT: the Available Integrations catalogue.

  • Cloud - a read-only role or service principal you create. See Cloud Exposure, AWS for the cross-account role, and Azure for the per-sync-type app registration permissions Azure needs.
  • Source control - a GitHub App installation, a GitLab OAuth authorization for one group, or an Azure DevOps service principal. See Source control.
  • Jira - an OAuth authorization.
  • EDR, appliances, IPAM, private cloud and directory - read-only API credentials from the vendor, or a domain account for Active Directory, and for anything inside your network an Integrator to reach it.

Active Integrations​

Everything you have connected, with its status and sync configuration, lives under Active Integrations. It is where you go to check an integration is still healthy, change which sync types are enabled, or re-authorize one whose credentials have expired.

Worth checking periodically rather than only when something looks wrong: an integration whose token has been revoked usually reports no error at all, because nothing is failing - the vendor has simply stopped answering.

Sync configuration​

An integration is not all-or-nothing. Each exposes sync types you enable individually, which is how you take a vendor's inventory without its log volume, or its alerts without its telemetry.

That granularity is worth using deliberately. Log-shaped sync types - traffic logs, CloudWatch, deep visibility - carry far more volume than inventory ones, and volume is the largest consumption dial in the platform. See Detection engineering.

Network-local targets​

Where a target has no public endpoint - an on-premises Jira, an internal source control server - the platform reaches it through an Integrator deployed inside your network.

When an integration stops working​

Almost always credentials:

  1. Expired or revoked credentials. OAuth grants get revoked, API keys rotate, service principal secrets expire. This usually presents as "data stopped updating" rather than an error, because nothing is failing - the vendor has stopped answering.
  2. Reduced permissions. A scope or role narrowed on the vendor side removes some sync types and leaves others working, so the integration looks healthy while part of it is not.
  3. Removed webhooks. For source control, a repository or organization policy change can remove the webhook registration. See Webhooks.

Reconnecting re-authorizes and re-registers, which fixes most of these.

API​

The endpoints behind this section are in the API reference.

Next​

  • AWS - the cross-account role AWS needs and connecting an organization
  • Azure - the app registration permissions each Azure sync type needs
  • Source control - GitHub, GitLab and Azure DevOps
  • API keys - for integrations you build yourself
  • Webhooks - the inbound endpoints vendors post to
  • Power Platform - the registration Power Platform inventory needs
  • Active Directory - reading a domain through an Integrator
  • Cloud Exposure - connecting a cloud account