Upgrade Core Operators and Core Instances
Use the information in this section to upgrade your Core Operators and Core Instances.You must use the same method for upgrading your Core Operators and Core Instances
that you used when you originally installed them.
Determine your upgrade method
The method you must use for upgrading your Core Operator and Core Instance depends on the method you used when you installed that Core Operator and Core Instance, which in turn depends on your environment.- Kubernetes cluster
- Linux
If you installed Telemetry Pipeline in a Kubernetes cluster and aren’t sure
whether you used Pipeline CLI or Helm during that installation process, run the
following kubectl command:The resulting output shows any labels attached to the
controller-manager
resource.- If the label
app.kubernetes.io/managed-by=Helmis present, that means you used Helm to install Telemetry Pipeline and should continue to use Helm for any subsequent upgrades to your Core Operator and Core Instance. - If the label
app.kubernetes.io/managed-by=Helmis not present, that means you used Pipeline CLI to install Telemetry Pipeline and should continue to use Pipeline CLI for any subsequent upgrades to your Core Operator and Core Instance.
Upgrade your Core Operator and Core Instance
After you determine your upgrade method, use the corresponding steps to upgrade your Core Operator and Core Instance to the latest available version.- Pipeline CLI
- Helm
If you deployed Telemetry Pipeline inside a Kubernetes cluster, Pipeline CLI must have
access to that cluster. If you deployed Telemetry Pipeline on Linux, you must
run Pipeline CLI from inside that Linux environment.
- Update Pipeline CLI. Because Core Operator and Core Instance versions are dependent on your Pipeline CLI version, you must use the latest version of Pipeline CLI to perform upgrades.
-
Run the following command to install the latest version of Core Operator, which
overwrites any previously installed versions:
Optionally, you can use the
--kube-namespaceflag to set the Core Operator’s namespace. -
Run the following command to update your associated Core Instance:
Replace
INSTANCEwith the name of the Core Instance you want to update.
Upgrade a pipeline
Regardless of what method you used to install and upgrade your Core Operator and Core Instance, you can use either the Telemetry Pipeline web interface or Pipeline CLI to upgrade pipelines.- Web interface
- Pipeline CLI
There are two possible methods for upgrading pipelines in the Telemetry Pipeline
web interface.Upgrade to the latest versionTo upgrade a pipeline to the latest version, follow these steps:Choose the agent image type and versionTo choose the agent image type and version
for your pipeline, follow these steps:
- Sign in to Telemetry Pipeline.
- Go to Core Instances, and then click the name of the pipeline you want to upgrade.
- If a pipeline isn’t using the latest agent version, an
Upgrade to <version> availablenotice is displayed on the pipeline details page. Next to this notice, click Upgrade to upgrade the pipeline to the latest version.
This upgrade method always retains a pipeline’s current
agent image type.
- Sign in to Telemetry Pipeline.
- Go to Core Instances, and then click the name of the pipeline you want to upgrade.
- Go to Advanced Settings.
- In the Pipeline Image section, select an option from the Version
drop-down menu. Version numbers with no suffix (for example,
1.2.3) use the Red Hat UBI agent image, and version numbers with the distroless suffix (for example,1.2.3-distroless) use the distroless agent image. - Click Apply to save your changes.
Upgrade multiple pipelines
To upgrade all of your pipelines in a given environment simultaneously, run the following command:VERSION with the latest version of the
pipeline agent. You can find this value in the Pipeline agent column of the
Telemetry Pipeline component versions table.
For example, if the latest version of the pipeline agent is 1.23.0, provide the
value 1.23.0 to use the Red Hat UBI agent image
or 1.23.0-distroless to use the distroless agent image.
For critical pipelines, Chronosphere recommends updating each pipeline individually
instead of using this script.