How to make a Health-Check curl call in Circle-Ci - docker

Question

I am trying to make a health check during my circleci workflow to see if the app compiles successfully. Despite the app building on circleci, curl is never able to find to the localhost port.
I have tried this locally on my machine and it works great.
Below is my workflow job:
health-check:
docker:
- image: image_name
steps:
- setup_remote_docker
- restore_cache:
key: image-cache-ci-{{ .Environment.CIRCLE_KEY }}
- run:
name: start
command: |
docker load < image.tar
docker run -d graphql-ci:$CIRCLE_KEY "./node_modules/nodemon/bin/nodemon.js index.js"
docker run graphql-ci:$CIRCLE_KEY curl http://localhost:4000/.well-known/apollo/server-health
The app successfully compiles. However, the second line always returns the following:
12354314532412342: Loading layer 1.803MB/1.803MB
Loaded image: graphql-ci:234523451325424
34523453153145235345234523452345234
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (7) Failed to connect to localhost port 4000: Connection refused
Exited with code exit status 7
I have also tried to use wait-on - wait-on http://localhost:4000/.well-known/apollo/server-health && curl http://localhost:4000/.well-known/apollo/server-health
I have also placed the curl called in a separate run step after the build one
- run:
name: check that the server is up
command: |
docker load < image.tar
docker run graphql-ci:$CIRCLE_KEY curl http://localhost:4000/.well-known/apollo/server-health
I have also tried to use docker exec but then it just tells me that it cannot find my docker container (and then specifies the docker container image hash).
If I remove the detach flag -d then it will show that it successfully started.
[nodemon] 2.0.2
[nodemon] to restart at any time, enter `rs`
[nodemon] watching dir(s): *.*
[nodemon] watching extensions: js,mjs,json
[nodemon] starting `node index.js`
The default PORT set is 4000, I add a flag to the docker run command to include --env PORT=4000 to just confirm it. Didn't affect anything but I know that it is running.
I also gave pm2 a try for the server since then the same terminal will be available by default. This will run but makes no difference.
Really just want to curl a sever I know is loaded in one circleci job.
Entrypoint
CMD [ "node", "index.js" ]
EDIT:
I got closer to getting a successful result.
scripts
"curl": "if [[ $(curl -o /dev/null -w '%{http_code}' http://localhost:4000/.well-known/apollo/server-health) == 200 ]]; then echo 1; else echo 0; fi",
"ci": "./node_modules/nodemon/bin/nodemon.js index.js",
"health": "start-server-and-test ci http-get://localhost:4000/.well-known/apollo/server-health 'npm run curl'",
curl returns 1 if successful (returns 200) and 0 otherwise.
I also installed a package called start-server-and-test which starts the server, waits for it to finish then run the final script. It is normally used for cypress but works here too.
This was the result
> #bespokemetrics/graphql# curl /usr/src/app
> if [[ $(curl -o /dev/null -w '%{http_code}' http://localhost:3000/.well-known/apollo/server-health) == 200 ]]; then echo 1; else echo 0; fi
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 17 100 17 0 0 8500 0 --:--:-- --:--:-- --:--:-- 8500
1 // this was the returned
npm ERR! code ELIFECYCLE
npm ERR! errno 143
npm ERR! package_name: `./node_modules/nodemon/bin/nodemon.js index.js`
npm ERR! Exit status 143
Why does it return an ERR despite succeeding?

Answer 1

I also created an npm script to container all of the code needed to trigger the test on 1 line.
- run:
name: Start and Curl Server
command: |
docker load < image.tar
docker run graphql-ci:$CIRCLE_SHA1 npm run health
"curl": "if [[ $(curl -o /dev/null -w '%{http_code}' http://localhost:4000/.well-known/apollo/server-health) == 200 ]]; then echo 'Success'; pkill node; else echo 'Failure'; fi"
"ci": "./node_modules/nodemon/bin/nodemon.js index.js"
"health": "start-server-and-test ci http-get://localhost:4000/.well-known/apollo/server-health 'npm run curl'"
I installed a package called start-server-and-test which allows you to start your server, wait for it to build, then run whatever command you want all on 1 line.
My curl command only returns the status code, if 200 then return a success message and kill the node process. That was the err that was returned after the success earlier because it wasn't closed properly.