Dynatrace’s Helm charts, located in different Github repositories are in a state of transition, and may be confusing to understand.
Some of our Helm charts can be found at github.com/Dynatrace/helm-charts. This directory includes both OneAgent Operator and Dynatrace Operator Helm charts. These charts are at least a few months old but remain available to protect any existing file references. One of these charts, (OneAgent Operator) is still used in certain marketplaces, such as Google Cloud, and Rancher, but is outdated and will be deprecated.
Active development is performed against the Dynatrace Operator Helm chart located at github.com/Dynatrace/dynatrace-operator/tree/{version}/config/helm. This deploys the Dynatrace Operator, which is under active development, and provides features such as Cloud-Native Full-Stack monitoring and containerized ActiveGate pods. The current version of this helm chart is “v0.6.0” as shown in the release page of the Github repository.
Default resource limits for the Dynatrace components are available. These limits can be set in the values.yaml file according to the sample Dynakube file located in the following subdirectory: github.com/Dynatrace/dynatrace-operator/blob/master/config/samples/ - please note that the sizing for the ActiveGate are large to protect memory consumption on the node. We have seen some high memory consumption for large clusters as the ActiveGate processes hundreds of thousands of events coming from the Kubernetes API Server. The size may vary, however, in most situations, and we plan on publishing better defaults. Results of internal sizing tests will be published sometime this summer.
One additional note about our plans for the currently active Helm chart.
The current version of the helm chart (for Dynatrace Operator) deploys all components and provides a values.yaml file for the Dyankube custom resource. The exception is the CRD, which is not deployed by Helm due to well-documented reasons discussed in the Helm community related to CRD migrations. If you’re interested in those issues, you can google “Github issues cert-manager CRD migrations” for plenty of rich discussion in the cert-manager github community.
To alleviate some of these CRD issues we are planning some refactoring of the chart this summer, likely around the v0.7.0 timeframe. Since the DynaKube is a relatively small file, we will offer a new values.yaml file that will allow you to configure the Operator, CSI Pod, and Webhook components, as well as reference an optional CRD file, which is currently deployed using kubectl.
This version (expected to be v0.7.0) will also allow us to release newer versions of the Operator on the GKE and Rancher marketplaces.
If you have questions about this, please let us know. The Dynatrace community website is a great resource for questions and allows all our customers to benefit from answers like these.
Kubernetes beatings will continue until morale improves
