Using plumber in a Rocker container

In this post I’ll go over my steps to create a docker image with plumber package to have a REST API endpoint on EC2 Amazon instance. This means that this endpoint will “live” and respond to requests from any tool of your choosing.

Initially, I’ve set up to do this in order to add Slack Slash command. However, I didn’t look into documentation first (RTFM!) and apparently slash commands can only return text. In my case I’ve created a ggplot2 plot that can’t be transferred. Oh, well :).

Creating a custom plumber image

Tutorial on how to use Docker with plumber is here.

  1. First step is obvious - docker pull trestletech/plumber.

  2. To run docker run --rm -p 8000:8000 trestletech/plumber you don’t need plumber on your local machine. However, there is a problem in examples since all end files were renamed to plumber.R. That means that if you want to run docker run --rm -p 8000:8000 trestletech/plumber /usr/local/lib/R/site-library/plumber/examples/04-mean-sum/meansum.R, you’ll be greeted with Error in plumber::plumb(commandArgs()[4]) : File does not exist: /usr/local/lib/R/site-library/plumber/examples/04-mean-sum/meansum.R. So, instead you want to write: sudo docker run --rm -p 8000:8000 trestletech/plumber /usr/local/lib/R/site-library/plumber/examples/04-mean-sum/plumber.R. I’ve opened an issue so at some point it’ll be fixed, hopefully.

  3. Also, there is no tail method there, so to test that everything is working, you should instead try in the terminal: curl localhost:8000/mean.

  4. Customizing Docker image is done in Dockerfile. In the end, my Dockerfile is looking like that:

FROM trestletech/plumber
MAINTAINER mishabalyasin <>
RUN R -e "install.packages('ggplot2')"
ADD plumber.R /app/
  1. With Dockerfile in the current directory I ran docker build -t plumber-heart . (file plumber.R should be in the same folder as Dockerfile). In my case, plumber.R was a file that drew a heart like this:
# code for the heart is from here -
#' @get /heart
#' @png
  heart.up <- function(x) {sqrt(1-(abs(x)-1)^2)} #Upper part of the heart
  heart.dw <- function(x) {acos(1-abs(x))-pi}    #Lower part of the heart
  #Plot of the heart
  p <- ggplot(data.frame(x=c(-2,2)), aes(x)) +
    stat_function(fun=heart.up, geom="line", aes(colour="heart.up")) +
    stat_function(fun=heart.dw, geom="line", aes(colour="heart.dw")) +
    scale_colour_manual("Function", values=c("red","red"), breaks=c("heart.up","heart.dw"))+
    labs(x = "", y = "") +
    guides(color = FALSE) +
  1. Finally, running the image is docker run --rm -p 8000:8000 plumber-heart /app/plumber.R. One small detail about using ggplot is that you need to explicitly print resulting plot. Otherwise you’ll get this error.

At this point you can test that everything is working by going to http://localhost:8000/heart

Sharing resulting image on Docker hub

  1. First, login to Docker hub: docker login --username=mishabalyasin

  2. Type docker images to find the ID of the image that was created in step 5 above.

  3. Assign tag docker tag <IMAGE_ID> mishabalyasin/plumber-heart:working.

  4. And, finally docker push mishabalyasin/plumber-heart.

Deploying in the cloud

Final step is to deploy it to the cloud since we want to use GET method with, e.g., curl.

  1. Create Amazon Linux instance. Free tier is available and enough for this type of endpoint.

  2. Install Docker as explained here.

  3. docker pull mishabalyasin/plumber-heart:working.

  4. Finally, we are ready to docker run --rm -d -p 8000:8000 <IMAGE_ID> /app/plumber.R (Notice that you need to use ID, not the name of the image).

Congratulations - you have your own API endpoint!


There aren't any comments yet. Be the first to comment!

Leave a comment

Thank you

Your comment has been submitted and will be published once it has been approved.



Your post has not been submitted. Please return to the form and make sure that all fields are entered. Thank You!