Skip to content

Audit overview

The audit log is the immutable record of everything that happened in your tenant. Every sign-in, every admin operation, every configuration change, every security event. Append-only; queryable in the console; exportable; streamable.

In short: actions that change state, fail to change state, or have security implications.

Some examples by family:

  • Authentication — sign-in success, sign-in failure, sign-out, session revocation, MFA challenge succeeded / failed.
  • User lifecycle — create, update, disable, enable, delete, password reset.
  • Admin operations — application config changes, federation connection edits, role assignments, flow / action changes, branding edits.
  • Security — brute-force detection, refresh-token reuse detection, breach incident matches, threat-feed hits.
  • Policy changes — authentication policy, password policy, MFA enforcement changes.

Not audited (would create noise without value):

  • Individual API calls authenticated against an already-issued token (those would multiply 100×).
  • Internal platform housekeeping (token rotation, cache eviction, etc.).
  • Things that affected no state (a sign-in attempt that didn't even reach the credential check, e.g., a malformed request).

Every entry has:

  • id — opaque event id.
  • type — closed taxonomy (user.signed_in, application.created, etc.). See Event schema.
  • occurred_at — ISO timestamp at platform precision.
  • actor — who did it. User / admin / system / application.
  • target — what was acted on. User / session / application / etc.
  • outcomesuccess / failure / denied / partial.
  • data — type-specific fields with the operation's details.
  • context — IP, user agent, fingerprint signal, request id, geolocation.

Audit → Read logs. A queryable table view. See Read the audit log.

For real-time monitoring, the Dashboard's recent-activity tile surfaces the last 20 admin-interesting events. For long-form investigation, the full audit surface is the answer.

Bulk-export audit entries for compliance evidence, vendor evaluation, or your own archive. See Export.

For long-term retention / centralised security monitoring, stream audit events to your own destination — webhook, Splunk, Datadog, Elastic. See Stream to a destination.

The platform retains audit entries for a configurable window:

  • Default: 365 days.
  • Configurable in Settings → Audit policy → Retention. Up to the maximum your plan allows.

After retention expires, entries are permanently dropped from the audit feed. If you need longer, stream to your own archive.

Common audit-use scenarios:

  • SOC 2 — auditors ask "show me all admin actions in <quarter>". Export with that filter.
  • GDPR / DPIA — "show me data-deletion events for this user". Filter on target_id = usr_X and event-types matching user.deleted or *.erasure_*.
  • Incident investigation — "what happened between 14:00 and 16:00 that triggered the alert?" Filter by time + outcome.
  • Internal review — "who changed the authentication policy in the last quarter?" Filter by event-type admin.policy_updated.
  • What's in a user's profile right now. The user record's current state is in the Users surface. Audit is the history of how it got there.
  • Whether a webhook receiver successfully processed an event. Audit records the platform-side fire; your receiver's processing is your code's concern. (Some platforms include receiver-confirm callbacks; IntelliAuth doesn't today.)

Audit is tenant-scoped. Events in production-cymmetri don't show up in staging-cymmetri. For cross-tenant audit (the platform admin view), the IntelliAuth admin console has its own audit surface — see CP admin audit.