Platform Audit Logging
Review administrative activity in Settings → Audit Log, and search, alert and build dashboards on platform audit events as OCSF API Activity.
Review the platform audit trail to find administrative and security-relevant actions. You can inspect sign-ins, rule changes, user and role changes, API key activity, and settings changes in Settings → Audit Log or Search.
- Settings → Audit Log: a filterable table of audit entries, with CSV/JSON export.
- Search: the same events as OCSF API Activity (
class_uid=6003) withmetadata.log_source=audit, so you can query them with nPL, chart them and write detections on them.
Audit Log page
Open Settings → Audit Log (/settings/audit; /settings/audit-log redirects there). You need the audit:view permission.
Filter by user, action, resource type, subsystem (source), success or failure, and time range. Users without the settings:system permission see only their own entries; administrators see everyone's.
Export
Export downloads CSV or JSON for the current filters, up to 100,000 rows (10,000 by default).
Every export is itself recorded in the audit trail.
API
| Method and path | Purpose |
|---|---|
GET /api/audit | List entries. Query parameters: user_id, action, resource_type, source, start_time, end_time, success, limit (max 1,000), offset |
POST /api/audit/export | Export with the same filters plus format (json or csv) and limit |
GET /api/audit/actions | Distinct action names, for filter pickers |
GET /api/audit/resource-types | Distinct resource types |
All four require audit:view.
curl -X POST -H "X-API-Key: $NANO_API_KEY" \
"https://<your-nano-host>/api/audit/export?format=csv&source=auth&success=false"Audit events in Search
Each audit entry is also written as an OCSF 1.9 API Activity event. Search them with:
metadata.log_source=auditAudit events are stored in ClickHouse. Security-critical events (sign-ins, permission and key changes, settings changes and similar) are also written to PostgreSQL first. If the ClickHouse write fails, a background worker retries it until ClickHouse accepts it, and the Audit Log page reads those rows (flagged as degraded) while ClickHouse can't answer.
Field reference
| OCSF field | Contents |
|---|---|
class_uid / class_name | 6003 / API Activity |
activity_id / activity_name | Derived from the action verb: 1 Create, 2 Read, 3 Update, 4 Delete, 99 Other (logins, denials, runs) |
metadata.log_source | audit |
metadata.product.name | nano |
metadata.uid | The audit event ID (stable across retries) |
api.operation | The nano action, for example login_failed, rule_deleted, apikey_created |
api.service.name | The subsystem: auth, detection, user, role, group, apikey, settings, … (see below) |
api.request.uid | The HTTP request ID, when known |
actor.user.name | Who acted (display name or email) |
actor.user.uid | The acting user's ID |
actor.user.email_addr | Set when the actor name is a plain email address |
actor.user.credential_uid | The API key ID, when the action was made with an API key |
src_endpoint.ip | Client IP address |
http_request.user_agent | Client user agent |
resources | The affected object: type, uid, name |
status / status_id | Success (1) or Failure (2) |
status_detail | Failure reason, on failed actions |
severity | Informational for successes, Low for failures |
message | Human-readable summary |
unmapped.* | The original nano fields: unmapped.source, unmapped.action, unmapped.resource_type, unmapped.resource_name, unmapped.api_key_name, unmapped.details, … |
See OCSF fields for the general schema.
Subsystems (api.service.name)
auth, detection, user, tuning, mitre, case, apikey, role, group, credentials, search, ingest, webhook, storage, identity, marketplace, dashboard, notebook, alert, query_library, log_source, source_config, enrichment, custom_enrichment, agent_enrichment, session, oidc, parser_repo, rule_repo, settings, risk, lookup, gdpr, audit, report, source_scope, artifact, hunt.
Case lifecycle actions are described in Case audit events.
Common actions (api.operation)
| Subsystem | Actions |
|---|---|
auth | login_success, login_failed, logout, token_refresh, password_reset_request, password_reset_complete, password_changed, oidc_login, auth_denied |
detection | rule_created, rule_updated, rule_deleted, rule_paused, rule_resumed, rule_promoted, rule_demoted, rule_duplicated, rules_bulk_updated |
user | user_created, user_updated, user_deleted, user_locked, user_unlocked, user_disabled, user_enabled, user_groups_updated |
role / group | role_created, role_updated, role_deleted, group_created, group_updated, group_deleted, group_roles_updated, group_member_added, group_member_removed |
apikey | apikey_created, apikey_updated, apikey_deleted, apikey_enabled, apikey_disabled, apikey_reset |
credentials | credential_created, credential_updated, credential_deleted, credential_rotated, credential_rolled_back |
source_scope | source_scope_restricted, source_scope_unrestricted, source_scope_granted, source_scope_revoked, source_visibility_group_set, source_visibility_member_set, source_visibility_bypass |
log_source | log_source_created, log_source_updated, log_source_deleted, log_source_published, log_source_reverted, log_source_deployed |
settings / tuning | system_settings_updated, settings_updated, proposal_approved, proposal_rejected, version_reverted |
The Audit Log page's action filter (GET /api/audit/actions) lists every action recorded in your deployment.
Example searches
Failed logins:
metadata.log_source=audit api.operation=login_failed
| stats count by actor.user.name, src_endpoint.ip
| sort -countEverything one person changed:
metadata.log_source=audit actor.user.name="alice@example.com" activity_id IN (1, 3, 4)
| table time, api.service.name, api.operation, messageFailed or denied operations, with the reason:
metadata.log_source=audit status=Failure
| table time, actor.user.name, api.operation, src_endpoint.ip, status_detailActivity by subsystem and action:
metadata.log_source=audit
| stats count by api.service.name, api.operation
| sort -countDaily admin activity trend:
metadata.log_source=audit
| timechart span=1d count by api.service.nameActions taken with an API key:
metadata.log_source=audit actor.user.credential_uid=*
| stats count by actor.user.name, api.operationDetections on audit events
Write detections on audit events like any other source. Some starting points:
| Detection | Query |
|---|---|
| Brute force | metadata.log_source=audit api.operation=login_failed | stats count by src_endpoint.ip | where count > 10 |
| Detection rule deleted | metadata.log_source=audit api.operation=rule_deleted |
| Role or group permissions changed | metadata.log_source=audit api.operation IN (role_updated, group_roles_updated) |
| New API key | metadata.log_source=audit api.operation=apikey_created |
| Source visibility changed | metadata.log_source=audit api.service.name=source_scope |
| Repeated failures by one user | metadata.log_source=audit status=Failure | stats count by actor.user.name | where count > 20 |
See Detections for scheduling and alerting.
Troubleshoot
Need help with these checks? Contact support with the request ID and time with timezone.
- An action is missing from Search but shows on the Audit Log page. A security-critical event whose ClickHouse write failed is retried in the background, and SIEM Health reports the backlog. Check the time range and try again shortly.
actor.user.nameis empty. System actions (scheduled jobs, platform runners) have no interactive user.- You only see your own entries. Viewing other users' entries needs
settings:systemas well asaudit:view.
What's next
Source Visibility
Hide events from specific log sources with a restricted-source registry and group grants.
Evidence Pack
Export compliance evidence for an audit period as JSON, CSV and PDF, with an optional mapping to SOC 2, ISO 27001, Essential Eight, ISM, NIS2 and DORA controls. Metadata only, every file hashed. Self-run, from v2.0.25.