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
Before I merge a PR I usually do a smoke test which involves:
Checking out the PR branch
Building the library
Deploying the example project to a sandbox AWS environment
Firing an event though and checking the logs to see if it worked
We can automate this process to make it quicker to merge PRs.
We already have a CI pipeline that runs via github actions and builds/tests the library. We just need to write an integration test step. I have already done a similar thing in one of my personal projects, so that might be a good reference.
The main things that the above script does that is nice to have is:
Has a trap that cleans up the infra if an error occurs
Generates a stage name from the git commit so that a fresh deployment is run every time
Some checks for tools/credentials
The hard part would be writing the part of the script that verifies if the deployment is working. My current thinking is:
Update the example project so that incoming events are dumped to an s3 bucket as JSON files
Fire an SNS event from the command line, which should then end up in a queue and then picked up by the lambda. Then poll the S3 bucket until the event file appears
Do a simple check on the event using jq
The text was updated successfully, but these errors were encountered:
Before I merge a PR I usually do a smoke test which involves:
We can automate this process to make it quicker to merge PRs.
We already have a CI pipeline that runs via github actions and builds/tests the library. We just need to write an integration test step. I have already done a similar thing in one of my personal projects, so that might be a good reference.
In my personal project, when the CI runs, it changes into the directory of the example project and runs the "integration-test" yarn task.
That then calls a bash script: https://github.com/NoxHarmonium/io-ts-serverless-handler/blob/master/packages/io-ts-serverless-handler-example/scripts/integration-test.sh
The main things that the above script does that is nice to have is:
The hard part would be writing the part of the script that verifies if the deployment is working. My current thinking is:
jq
The text was updated successfully, but these errors were encountered: