Skip to main content

High Availability

Updated Feb 05, 2023 ·

Single Traefik Instance​

Running a single Traefik instance is simple, but it has its risks. If Traefik crashes or stops, traffic won’t reach your containers.

  • Containers keep running, but new requests get blocked.
  • Issues like disk space filling up can cause failures.
  • Updating Traefik may lead to downtime.
  • During these moments, your service becomes unavailable.

Traefik Cluster​

To avoid downtime, use multiple Traefik instances working together in a cluster.

  • Run several Traefik instances at the same time.
  • Deploy each instance on a different server.
  • If one server fails, the others keep handling traffic.

A typical high-availability setup includes:

  • Multiple Traefik instances:

    • At least three, on different hosts.
    • Keeps traffic flowing if one node fails.

  • Virtual or Elastic IP:

    • One IP address for all traffic.
    • Automatically routes to healthy instances.

  • Common storage:

    • Shared place for storing SSL certificates.
    • Also holds configuration files, keeping instances in sync

Common Storage​

Traefik instances need to share certificates and configurations using common storage.

  • They elect a leader to handle updates like renewing certificates.
  • This keeps everything in sync and working well.

How HA Traefik Works​

Setting up HA Traefik involves:

  • Incoming traffic goes to the virtual IP.
  • The virtual IP routes requests to any healthy Traefik instance.
  • Each Traefik reads and writes to the shared storage.
  • Leader election makes sure updates are handled properly.

This setup helps avoid failures and downtime, keeping your services available all the time.