forbidden error installing ambassador #596
Comments
agrski
added a commit
that referenced
this issue
Dec 2, 2022
* Add CPU request parameterisation to MLServer & Triton ServerConfigs * Add default CPU requests to Helm values for ServerConfigs * Add resource limits in ServerConfig Helm templates * Remove trailing whitespace in Kustomize patches for Helm * Ensure all components in Helm chart have CPU & memory resources parameterised Also set default CPU & memory requests & limits. * Use consistent units & formatting for resource requests * Add Helm parameterisation for scheduler storage volume * Regenerate components Helm chart * Fix incorrect parameterisation of scheduler storage requests The right Helm variable was specified in the wrong place. * Simplify container resource config in Helm charts Use just CPU, memory, and storage in user-facing values and set requests & limits in the underlying template. This change cover the Kustomize patches and Helm values template. This is embedding best practices (set memory requests to limits, don't set CPU limits) but has the added benefit of being simpler for a user to interact with. * Regenerate components Helm chart * Fix scheduler storage parameterisation in Kustomize patch * Remove hard-coded CPU limit for operator * Reduce dataflow engine CPU requirements in Helm chart This is brought in line with the model & pipeline gateways. One might reasonably expect them all to have similar levels of work to do. * Use consistent units for CPU requirements in Helm values
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We're using v0.40.2 with the latest helm chart. But the chart is now tailored to v0.70.1 and the pods going into CrashLoopBackOff with a permissions error (forbidden) about watching configmaps
The text was updated successfully, but these errors were encountered: