App Engine Flex + Docker + R +Plumber does not work only when I use 'AI Platform Deep Learning Containers'

[ashishsr123]

I want to serve R model using App Engine. I am following this example R with app engine, but stuck. everything works when I do not use 'AI Platform Deep Learning Containers' ,details below.

Please refer to my code app.yaml

service: caret-custom
runtime: custom
env: flex

readiness_check:
  app_start_timeout_sec: 1800

resources:
  cpu: 1
  memory_gb: 0.5
  disk_size_gb: 10

Dockerfile

FROM gcr.io/deeplearning-platform-release/r-cpu  
#FROM gcr.io/gcer-public/plumber-appengine  <--  code works using this container


RUN R -e "install.packages(c('caret', 'xgboost', 'rjson', 'plumber','googleCloudStorageR','googleAuthR'), repos='http://cran.rstudio.com/')"

WORKDIR /payload/
COPY [".", "./"]

EXPOSE 8080
ENTRYPOINT ["R", "-e", "pr <- plumber::plumb(commandArgs()[4]); pr$run(host='0.0.0.0', port=8080,swagger=TRUE)"]
CMD ["main.R"]

main.R

print("in rest controller program")
#* @get /demoR
get_predict_length <- function(){
      
  dataset <- iris
  
  # create the model
  model <- lm(Petal.Length ~ Petal.Width, data = dataset)
  petal_width = "0.4"
  
  #petal_width = '0.4'
  # convert the input to a number
  petal_width <- as.numeric(petal_width)
  
  #create the prediction data frame
  prediction_data <- data.frame(Petal.Width=petal_width)
  
  # create the prediction
  predict(model,prediction_data)
}

I deploy using 'gcloud app deploy and its fails with below error ERROR: (gcloud.app.deploy) Error Response: [4] Your deployment has failed to become healthy in the allotted time and therefore was rolled back. If you believe this was an error, try adjusting the 'app_start_timeout_sec' setting in the 'readiness_check' section.

I increased 'app_start_timeout_sec' gradually from 300 to 1800 and now I get below error: Updating service [caret1-custom] (this may take several minutes)...failed.
ERROR: (gcloud.app.deploy) Operation [apps/my-project-id/operations/7db0da8f-213b-476f-a6ba-9c8102b21cd5] timed out. This operation may still be underway.

Issue: If I use 'FROM gcr.io/gcer-public/plumber-appengine' in place of google provide container 'gcr.io/deeplearning-platform-release/r-cpu', everything is successful.

Please advice, as I am plan to use google provided containers only.

Ashishsr123