Closed openjck closed 5 years ago
Also, it should be noted that @pdehaan did do some eary load testing of the whole stack a few months ago in this issue:
I don't think we will need to loadtest the transposer on Heroku if we're using autoscaling. We can save that effort for when/if OPs migrates it to AWS/GCP/whatever.
Okay great. And in regard to your second sentence, we may soon rewrite ensemble-transposer as a Lambda function that writes data to S3, so we may not even need to do that ultimately.
For launch, we will be running ensemble-transposer on Heroku with autoscaling enabled. DataOps will be monitoring uptime and will address any issues.
@pdehaan Given that we will be using autoscaling, do you think it's worthwhile to load test ensemble-transposer before launch? If it would be useful, I can temporarily enable autoscaling and serve up real data for the duration of the tests.