Securing and operating the cluster

Brightlane's cluster has run open and unwatched so far: no logins, no permissions, and nobody told when something breaks. This module makes it production-ready: encrypted connections and a login per service, permissions and quotas, metrics and alerts that point at the right team, and changes made while traffic keeps flowing.

By the end of this module you can: - set up a SASL_SSL listener with TLS and SCRAM, and connect Python clients to it; - tell authentication from authorization errors, and read the client's message for each failure; - give each service least-privilege ACLs, and limit a client with a quota; - collect broker, lag and client metrics in Prometheus, and write alert rules with runbooks; - manage topics as code, restart and upgrade brokers without downtime, and move replicas with a throttle.

Lessons

  1. Encryption and authentication
  2. Authorization and quotas
  3. Observability: metrics, lag and alerts
  4. Day-2 operations: change without downtime

Then practise with the exercises and check yourself with the module quiz.