Skip to content

Audit Logs

To help comply with various data protection regulations, Welkin comes built-in with audit logs, which can be accessed via OpenSearch Dashboard.

What are audit logs?

In brief, audit logs are lines answering "who did what and when?".

Why are audit logs important?

Audit logs help both with proactive and reactive security:

  • Regular audit log reviews give you a chance to catch an attacker before they succeed.
  • After-the-fact, audit logs allow you to gather evidence for forensics and assess the extend of the damage caused by an attacker.

What audit logs are included?

Welkin follow a risk-based approach to audit logs. We enable audit logs on all APIs which can be used to compromise data security. Then we filter high-volume low-risk audit logs. Don't forget that, at the end of the day, logs are only as useful as someone looks at them. See log review for details.

Specifically, the following audit logs are configured by default:

  • Kubernetes API audit logs;
  • SSH access logs.

Further audit logs can be configured on a case-by-case basis, as described below.

Kubernetes API Audit Logs

The audit logs are stored in the kubeaudit* index pattern. The audit logs cover calls to the Kubernetes API, specifically who did what and when on which Kubernetes cluster.

Thanks to integration with your Identity Provider (IdP), if who is a person, their email address will be shown. If who is a system -- e.g., a CI/CD pipeline -- the name of the ServiceAccount is recorded.

Your change management or incident management process should ensure that you also cover why.

Both users (Application Developers) and administrators will show in the audit log. The former will change resources related to their application, whereas the latter will change Welkin system components.

The exact configuration of the Kubernetes audit logs can be found here.

To view the audit logs for a specific user:

  1. Open the Audit user dashboard in OpenSearch;
  2. Under User selector add the name of the user you want to audit (e.g admin@example.com);
  3. Apply changes;

Example of Audit Logs

SSH Access Logs

Welkin also captures highly privileged SSH access to the worker Nodes in the authlog* index pattern. Only administrators should have such access.

Example of SSH Access Logs

Note

This section helps you implement ISO 27001, specifically:

  • A.9.2.1 User Registration and Deregistration

Many data protection regulation will require you to individually identify administrators, hence individual SSH keys. This allows you to individually identify administrators in the SSH access log.

Audit Logs for Additional Services

The Kubernetes Audit Logs capture user access to additional services, i.e., kubectl exec or kubectl port-forward commands. Additional services usually do not have audit logging enabled, since that generates a lot of log entries. Too often the extra bandwidth, storage capacity, performance loss comes with little benefit to data security.

Prefer audit logs in your application to capture audit-worthy events, such as login, logout, patient record access, patient record change, etc. Resist the temptation to enable audit logging too "low" in the stack. Messages like "Redis client connected" are plenty and add little value to your data protection posture.

Out of all additional services, audit logging for the database makes the most sense. It can be enabled via PGAudit. Make sure you discuss your auditing requirements with the service-specific administrator, to ensure you find the best risk-reduction-to-implementation-cost trade-off. Typically, you want to discuss:

  • which databases and tables are audited: e.g., audit app.users, but not app.emailsSent;
  • what operations are audited: e.g., audit INSERT/UPDATE/DELETE, but not SELECT;
  • by which users: e.g., audit person access, but not application access.

Further Reading