Turn on the security modules
The security modules look for weaknesses in your SQL Server instances, your identities and your Azure subscriptions, and add findings from Microsoft Defender for Cloud and Wazuh. Every check only reads; a finding comes with its evidence and a suggested action for a person to follow, and nothing is ever changed for you. This guide shows how to switch the modules on, give them read-only access, weight the score and accept a risk.
The security modules and the security score are part of the Enterprise plan, and an add-on in the Pro plan.
Before you begin
- To switch the modules on, you need the Super admin role.
- To add resources and credentials, and to set the score weights, you need the Entity admin role on the entity.
- To accept a risk, you need the Editor role or higher on the finding's entity.
- Read Security posture for what a finding is, how the score is calculated and what each check detects.
Switch the modules on
-
Open Settings › Instance.
-
Under Features, in the Security group, switch on what you need:
Feature Effect securityThe Security posture page, the findings and the score, with the database, identity and cloud modules security.recommendationsDefender for Cloud recommendations (medium and low assessments) are shown, counted in the score and listed in the inbox. When off, only high-severity assessments remain, as findings security.defenderCollection of Defender for Cloud assessments. Off by default security.wazuhCollection of Wazuh alerts and vulnerabilities. Off by default -
Optionally switch on
collab.jiraso findings can be sent to Jira.
The other three switches require security. Switching Defender or Wazuh off stops collecting; findings already raised stay open until they are resolved.
Add the checks to existing resources
Resources you add from now on get the security checks of their type at once. Resources that existed before you switched the modules on need the missing checks added once, through the API:
curl -s "${H[@]}" -X POST "$BASE/resources/$SQL01/checks" -d '{"check_ids": []}' \
| jq -r '.[].check_id' | grep security
An empty list adds every missing default check of the resource's type. Discovered databases and availability groups get theirs at the next Discover of their instance. $BASE and H are set as in Quickstart: add a SQL Server instance with the API.
Give the modules read-only access
Database
The SQL Server security checks use the instance's monitoring login. It needs VIEW SERVER STATE, VIEW ANY DEFINITION, CONNECT ANY DATABASE and SELECT on msdb.dbo.backupset, and nothing else. See Store a credential for the collector.
Identity
- Add a resource of type Entra ID tenant / domain services with the tenant's
tenant_id, under the entity it belongs to, for example Ciguapa Inc.. - Give it read access to Microsoft Graph, in one of two ways:
- An app registration with the Microsoft Graph application permissions
RoleManagement.Read.Directory,AuditLog.Read.AllandIdentityRiskEvent.Read.All, all read-only, with admin consent. Store it as a credential of kind Azure service principal: the client ID, the client secret and thetenant_id. Choose it as the resource's credential. - The same permissions given to the managed identity of the Excubius server.
- An app registration with the Microsoft Graph application permissions
- For the Active Directory checks, provide a read-only domain account (credential kind Windows account) on a host that has the Active Directory PowerShell module, reached over WinRM.
Registration details for multifactor authentication need Microsoft Entra ID P1 or P2; full risk detections need P2.
Cloud
The Azure checks read through Azure Resource Graph and Defender for Cloud, with either the server's managed identity or an app registration stored as an Azure service principal credential:
- Reader on each subscription, for the Resource Graph checks.
- Security Reader on each subscription, for Defender for Cloud.
With an app registration, the credential's subscriptions attribute, set through the API, limits which subscriptions are read; otherwise the server's AZURE_SUBSCRIPTION_IDS setting applies. See Request the Azure server for the roles and their IDs.
The Resource Graph checks look at the resource's resource_id: a resource, a resource group or a whole subscription, such as /subscriptions/<id>.
Defender for Cloud and Wazuh
Defender for Cloud assessments, Azure role assignment changes, Wazuh findings and the platform's own image scan run on a resource of type Monitoring platform itself. If you do not have one yet, add it with Resources › Add resource, for example under the root entity. Findings are then attached to the resource they concern: Defender's to the inventory resource whose resource_id matches, Wazuh's to the host whose name, host name, IP address or machine name matches the agent.
For Wazuh:
-
Create a read-only Wazuh user with read access to the
wazuh-alerts-*andwazuh-states-vulnerabilities-*indices and, optionally,agents:readon the manager API. -
Store it in Settings › Credentials as kind HTTP user and password. If the indexer's certificate is not one the server trusts, create the credential through the API instead, with
"attributes": {"verify_tls": false}; the console has no field for it. -
Point the check at your Wazuh indexer and give it the credential, through the API:
curl -s "${H[@]}" -X PATCH "$BASE/resources/$PLATFORM/checks/security_cloud.wazuh_alerts" -d '{"params": {"indexer_url": "https://wazuh.areito.example:9200","api_url": "https://wazuh.areito.example:55000"},"credential_id": "'"$WAZUH_CRED"'"}'
The Wazuh client sends only searches and the manager's authentication call; anything else is refused in code. Excubius never deploys Wazuh agents; your own tooling does.
Weight the areas of the score
The overall score averages the assessed areas, identity, database and cloud, with a weight of 1 each by default. An entity can set its own weights, from 0 to 100, and every entity below inherits them. There is no screen for it in this release; use the API:
# Ciguapa cares most about its databases
curl -s "${H[@]}" -X PUT "$BASE/entities/$CIGUAPA/security/score-weights" \
-d '{"identity": 1, "database": 3, "cloud": 1}'
# back to the inherited weights
curl -s "${H[@]}" -X PUT "$BASE/entities/$CIGUAPA/security/score-weights" -d '{}'
GET /entities/{id}/security/score-weights shows the entity's own weights and the effective ones.
Accept a risk
When a finding is known and acceptable for a while, for example xp_cmdshell enabled for a legacy job until it is replaced, accept the risk instead of leaving it open.
- Open Security posture and select the finding.
- Select Accept risk….
- Write the Reason: why the risk is acceptable and what mitigates it, from 3 to 2,000 characters.
- Under Accepted for, choose a number of days, or Until a date. The end must be within a year.
- Select Accept risk.
The finding leaves the open list and stops counting in the score. It keeps being checked, and it reopens by itself when the acceptance ends, or earlier if its severity rises. The decision is kept with your name, the date, the end and how it ended.
To end an acceptance early, open the finding and select Reopen.
Some findings are changes rather than conditions, such as a new member of sysadmin or a new Owner role assignment. They cannot be accepted: review the change, then resolve the finding by hand.
Post a scan of the platform's images
The check security_cloud.platform_images reports critical vulnerabilities in Excubius's own container images. Produce a report with trivy image --format json <image> and post it to POST /api/v1/security/platform-report; this needs the Super admin role. The latest report of each image counts.