UPGRADING.md 2.1 KB
Upgrade guidelines
To 0.24
This release uses the kubernetes attributes processor
to add kubernetes metadata as resource attributes. If you override the processors array in your config, you will
need to add the k8s attributes processor manually to restore service.instance.id
resource attribute.
To 0.23
The Prometheus sub-chart dependency made updates to pod labels. You may need to
use the --force option with your Helm upgrade command, or delete the release
and re-install it.
To 0.22
This release moves to using the connectors functionality in the OpenTelemetry
Collector. The spanmetrics processor has been moved to use connectors
which results in an additional required exporter in the traces pipeline.
Existing releases that override exporters in the traces pipeline, will
need to add spanmetrics to the list of exporters before upgrading. The
OpenTelemetry Collector will fail to start otherwise.
To 0.21
The deployment labelSelector app.kubernetes.io/name has been renamed to
individual workload naming. If you upgrade it from charts <= 0.20, you
will have to delete all existing opentelemetry-demo deployments before running
helm upgrade command.
To 0.20
The observability.<sub chart>.enabled parameters have been moved to an
enabled parameter within the sub chart itself. If you had changes to these
parameters, you will need to update your changes to work with the new structure.
To 0.18
The serviceType and servicePort parameters have been moved under a service
parameter with names of type and port respectively. If you had changes to
these parameters for any demo component, you will need to update your changes
to work with the new structure for the service parameter.
To 0.13
Jaeger was moved to a Helm sub-chart instead of a local chart deployment. If you
had changes specified to the observability.jaeger parameter, those changes
will need to be re-implemented as sub-chart parameters under the top level
jaeger parameter instead.