You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After upgrading our operational version of datadog via helm in our test environment, I noticed that when trying to update the dd-dotnet-trace library version to a newer version was failing. Describing the pod image showed that the agent did not have access to the Azure ACR repository: datadoghq.azurecr.io
After reverting back to the dd-dotnet-trace library version that the cluster (AKS v1.29.15) had cached, I found that in the chart template, if you set your location/site to us3.datadoghq.com , the new default registry to pull new artifacts from is set to datadoghq.azurecr.io per lines 289-305 in the _helpers.tpl file, as shown below:
{{/*
Return the proper registry based on datadog.site (requires .Values to be passed as .)
*/}}
{{- define "registry" -}}
{{- if .registry -}}
{{- .registry -}}
{{- else if eq .datadog.site "datadoghq.eu" -}}
eu.gcr.io/datadoghq
{{- else if eq .datadog.site "ddog-gov.com" -}}
public.ecr.aws/datadog
{{- else if eq .datadog.site "ap1.datadoghq.com" -}}
asia.gcr.io/datadoghq
{{- else if eq .datadog.site "us3.datadoghq.com" -}}
datadoghq.azurecr.io
{{- else -}}
gcr.io/datadoghq
{{- end -}}
{{- end -}}
Unfortunately, the artifacts at datadoghq.azurecr.io are either not available, or do not have permissions set correctly, and this caused deployments that needed to get newer artifacts to fail deployment.
Temporary Solution was to explicitly set registry: gcr.io/datadoghq in the Values.yaml file and re-deploy.
The text was updated successfully, but these errors were encountered:
We default to Azure ACR on US3 to reduce transfer costs, notably for the Agent images. Nonetheless, as this repository is new, it could be missing some images. As you found out, you can override the registry, and you can even do so solely at the cluster Agent admission controller level to keep pulling Agent/cluster Agent images from ACR while pulling libraries from a different registry:
After upgrading our operational version of datadog via helm in our test environment, I noticed that when trying to update the dd-dotnet-trace library version to a newer version was failing. Describing the pod image showed that the agent did not have access to the Azure ACR repository:
datadoghq.azurecr.io
After reverting back to the dd-dotnet-trace library version that the cluster (AKS v1.29.15) had cached, I found that in the chart template, if you set your location/site to
us3.datadoghq.com
, the new default registry to pull new artifacts from is set todatadoghq.azurecr.io
per lines 289-305 in the _helpers.tpl file, as shown below:Unfortunately, the artifacts at datadoghq.azurecr.io are either not available, or do not have permissions set correctly, and this caused deployments that needed to get newer artifacts to fail deployment.
Temporary Solution was to explicitly set
registry: gcr.io/datadoghq
in the Values.yaml file and re-deploy.The text was updated successfully, but these errors were encountered: