Migrating away from the Linkerd-jaeger extension

Starting with the Linkerd 2.19 release, the Linkerd-jaeger extension will no longer receive updates. It will continue to work with the Linkerd 2.19 release and associated edge releases, but the extension is now unmaintained and may stop working in a future Linkerd release.

Migrating away from the extension

We recommend migrating away from the Linkerd-jaeger extension entirely and install dedicated tracing infrastructure on your cluster instead, e.g. by following our guide to distributed tracing with Linkerd.

Once you have the tracing infrastructure inside your cluster that is not managed by the extension, you can remove it safely from your cluster by running:

kubectl delete ns linkerd-jaeger

Then, restart all meshed pods.

Using the Linkerd-jaeger trace collector

If you already rely on the tracing collector and Jaeger instance installed by the extension, you may continue to use it as a temporary measure. Note that the tracing collector and Jaeger used by the extension are very old versions, but should continue to work with newer versions of Linkerd for the foreseeable future.

First, in your Linkerd-jaeger installation, set the following values:

webhook:
  enabled: false

Note

The most recent versions of the Linkerd CLI no longer have the linkerd jaeger commands. This guide assumes that you manage your Linkerd-jaeger installation through Helm or another similar mechanism. If necessary, you can use an earlier version of the CLI with these commands.

Then, in your Linkerd control plane installation, set the following values:

proxy:
  tracing:
    enabled: true
    collector:
      endpoint: collector.linkerd-jaeger:4317
      meshIdentity:
        serviceAccountName: collector
        namespace: linkerd-jaeger

Note

This assumes you have made no changes to the collector configuration in your Linkerd-jaeger installation. If you have changed the endpoint or already brought your own tracing collector, you will have to update the endpoint as well as the serviceAccountName and namespace in the meshIdentity in the above config. See our guide on distributed tracing for more information on how to set these values.

Finally, restart all meshed pods.