Merge pull request #91296 from frcai/patch-1

Update how-to-troubleshoot-online-endpoints.md

Author: Court72

articles/machine-learning/how-to-troubleshoot-online-endpoints.md

@@ -121,6 +121,7 @@ There are three supported tracing headers:
 
 Below is a list of common deployment errors that are reported as part of the deployment operation status.
 
+* [ImageBuildFailure](#error-imagebuildfailure)
 * [OutOfQuota](#error-outofquota)
 * [OutOfCapacity](#error-outofcapacity)
 * [BadArgument](#error-badargument)
@@ -129,6 +130,12 @@ Below is a list of common deployment errors that are reported as part of the dep
 * [OperationCancelled](#error-operationcancelled)
 * [InternalServerError](#error-internalservererror)
 
+### ERROR: ImageBuildFailure
+
+- Check the build log to diagnose common image build failure problems. The build log path is included in the failure message: 'The build log is available in the workspace blob store "your-blob-storage" under the path "/azureml/ImageLogs/your-image-id/build.log"'.
+- No obvious error was found in the build log and the last line is 'Installing pip dependencies: ...working...'. Pinning version dependencies in your conda file could fix this issue.
+- Use [local deployment](#deploy-locally) to test and debug your models locally before deploying in the cloud.
+
 ### ERROR: OutOfQuota
 
 Below is a list of common resources that might run out of quota when using Azure services: