Audit log
Overview
Section titled “Overview”Govern → Records → Audit log (/compliance/audit-log) is the organisation’s record of what people, API keys and Shield360 itself did. It holds three kinds of record:
| Kind | Tab | Recorded when |
|---|---|---|
| Change | Changes | Someone creates, changes or deletes something: a setting, a policy, a role, a key, a case |
| Access | Data access | Someone reads or exports sensitive data: a person’s or computer’s activity, a request’s prompt and response, guard events, sensitive-data findings, the audit log itself |
| Sign-in | Sign-ins | Someone signs in, fails to sign in, signs out, or is refused |
The log is append-only. The database refuses to change or delete an audit record, and records are kept for as long as the organisation exists. They are removed only when the organisation itself is deleted.
Who can see it
Section titled “Who can see it”| Permission | Allows | Held by |
|---|---|---|
compliance.audit.view (Read the audit log) | Open the audit log | Owners, administrators, compliance officers, the Security role and Auditors |
compliance.audit.export (Export the audit log) | Download it as CSV | Owners, administrators, compliance officers and Auditors |
An Auditor sees only the records about the one environment they were invited to, and only until their access ends. Everyone else with the permission sees the whole organisation’s log.
Reading the audit log and exporting it are recorded in the audit log too, with the filters used.
Find a record
Section titled “Find a record”Use the time filter at the top right of the page. Every count and list on the page covers that range.
Select Everything, Changes, Data access or Sign-ins.
Type in Search actions, paths, people to match the action name, the request path or the target. Select a name in the Who column to show only that person’s records (By …), or a target in the About column to show only records about it (About …). Select a chip to remove it.
Select a row to see the request (method, path and HTTP status), the actor type, the record ID and the record’s details.
The tiles at the top count Records, Changes, Data access, Sign-ins, Failed sign-ins and Denied (refused for lack of access) in the range. Most active lists the ten people with the most records.
What a record shows
Section titled “What a record shows”| Column | Meaning |
|---|---|
| When | When it happened |
| Who | The person’s email; API key "<name>" for a request made with an API key; System for Shield360’s own jobs |
| What | The action, in plain words where Shield360 has a label for it, for example Changed a member’s role |
| About | What it acted on, such as a person, a computer, a role or a key |
| Outcome | success; denied when the request was refused (401 or 403); failed for any other error |
| From | The client’s IP address, as seen by Shield360’s trusted proxy |
When a managed-service provider’s staff member acts under a grant, the record also says which provider they acted for.
The details never hold secrets. Fields named like a password, token, secret, API key, authorization header, cookie, credential or private key are dropped before the record is written. Text values are cut to 256 characters, and at most 20 fields are kept.
The list shows the latest 300 records. The export has every record that matches.
Export the audit log
Section titled “Export the audit log”Select Export (CSV) to download the records that match the current time range and filters as a CSV file named shield360-audit-log_<from>_<to>.csv, oldest first, up to 100,000 records per file. Its columns are Event ID, Time (UTC), Category, Action, Outcome, Actor type, Actor, Target type, Target, Method, Path, Status, IP address and Details (as JSON).
To keep the log in your own tools as it is written, stream it to a SIEM. See SIEM & security tools. Audit-log counts are also part of the evidence package.
What is recorded
Section titled “What is recorded”Every change made through Shield360’s API is recorded. A change with no label of its own is named from its path and verb, for example api-key.create or fleet-hub.config.create. The main actions, grouped:
| Area | Examples of actions |
|---|---|
| Sign-in | auth.sign-in (success or failed, with the reason, such as an unverified email or SSO being required), auth.sign-out, agent_oauth.authorize, agent_oauth.token |
| Members, roles and identity | members.remove, members.role_change, organisation.member.add, organisation.member.api_keys.revoke, organisation.project_role.update, auditor.invite, organisation.sso.update, organisation.sso.domain.verify, organisation.scim.token.create, scim.user.create |
| Keys and secrets | api-key.create, api-key.delete, vault.secrets.read (secrets read with an API key), vault.external.read |
| Policies and change control | Enforcement-policy saves and reverts (compliance.enforcement_policy.*), approving and rejecting policy changes, change-control.settings.update, change-control.freeze.start and .end, transports (transport.create, transport.approve, transport.release, transport.rollback) |
| Collection and data | collector.policy.update, collector.ai_endpoints.update, data_retention.update, data_classification.update, organisation.export.request, organisation.export.download, person-data.erase, person-data.export.request |
| Reading sensitive data | telemetry.span.read, visibility.person.read, visibility.computer.read, visibility.drilldown.read, compliance.data-protection.read, compliance.data-protection.event.read, compliance.data-protection.export, sensitive-data.findings.read, coding-agents.session.read, ai-accounts.identifier.reveal |
| Compliance records | compliance.audit-log.read, compliance.audit-log.export, risk.create, risk.update, findings.*, impact_assessments.*, intake.*, controls.soa.update, compliance.ediscovery.search, compliance.ediscovery.export |
| Agents | agent_authorization.profile.update, agent_authorization.grant.save, agent_authorization.kill, agent_authorization.elevation.approve, agent_authorization.mandate.create, data_gateway.authorization.create |
| Alerts, AI analyst and spend | alerting.rules.create, alerting.rules.update, analyst.settings.update, budgets.create, finance.export, pricing.catalog.refresh |
| Licensing and modules | organisation.modules.update, license.issue, license.revoke, managed.grant.create, managed.grant.revoke, managed.provider.access |
Related
Section titled “Related”Four-eyes approval of policy and detection-rule changes
Evidence package and auditor accessGive an external auditor read-only access to one environment
SIEM & security toolsStream audit and sign-in events to your SIEM
API keysRequests made with a key are recorded under the key’s name