Attack protection

The attack protection policy lets you customize your org's response to bot activity in your sign-in, sign-up, and password reset flows.

Early Access release

Attack protection builds the standard bot protection functionality into the Okta policy framework so that you can customize activity thresholds and responses for each of your authentication flows. Automated attacks pose a significant threat to organizations. Attackers use readily available tools and cloud services to gain access to user accounts, distributing their attacks across many different IP addresses to avoid detection by security measures that focus on IP-based reputation. Okta's attack protection feature detects and prevents these automated attacks before they can compromise user credentials. It provides you with ongoing, clear visibility of bot activity in your org and the remediation steps that were taken.

How it works

The attack protection policy uses a machine learning-based model that uses both stateful and stateless inputs to determine whether a request comes from a bot. If it does, you can enforce an Okta Challenge, which doesn't require user input.

Like other Okta policies, the attack protection policy contains a catch-all rule that logs all bot activity but doesn't enforce Okta Challenge. The rules that you add to the policy let you customize different thresholds and actions for each flow.

The Bot Likeliness thresholds can affect the number of false positives or false negatives.

  • High: Attack protection triggers an event only when there's a high chance of bot activity. This may result in more false negatives (like bots not being prompted for Okta Challenge).

  • Medium and above: Attack protection triggers an event when there's a moderate chance of bot activity.

  • Low and above: Attack protection triggers an event when there's a small chance of bot activity. This may result in more false positives (like humans being prompted for Okta Challenge).

  • Any: Attack protection triggers an event whenever there's a chance of bot activity.

The Actions are your org's response when the indicated level of bot activity is detected.

  • No further action: Bot detections are logged, but no additional remediation actions are enforced by the policy.

  • Okta Challenge: Logging with enforcement of Okta Challenge. Third-party CAPTCHA is overridden for your selected flow, but it may still be enforced on unprotected flows. You can disable this in Security > General.

Based on your chosen detection threshold, bot traffic is logged in the System Log.

Migration

When you enable the attack protection policy, your bot protection settings are automatically migrated.

  • Each flow that you configured for Okta Challenge in bot protection is migrated to its own rule in the attack protection policy. The bot likeliness threshold that you previously configured is retained, but the label is now Bot Likeliness.
  • Each flow that you configured for No further action in bot protection is migrated to its own rule in the attack protection policy with the same setting, but now your third-party CAPTCHA settings are re-applied. CAPTCHA is overridden only by selecting Okta Challenge for a flow. You can change your org-level CAPTCHA settings in Security > General.

User experience

Changes in the user experience are minimal.

Attack protection starts when user credentials are submitted for authentication. Depending on the flow, that's when the user clicks Next, Sign up, or Change password. If your policy is set to Monitoring, users don't notice any changes. If it's set to Enforced, users are presented with a brief "Verifying" message. Verified users continue to their dashboard, and requests that fail the verification are redirected to a 403 error page.