Alyx production docker fails to respond - change recipe for production Alyx

oliche commented 9 months ago

We currently deploy docker on an EC2 instance.

Recently the docker container is unresponsive. The logs are saved internally to the docker and haven't been copied since July. The alyx instance needs rebooting several times a week.

There are 2 issues:

docker versions and running instances are not managed by AWS as they could be in Apps or Beanstalk
all of the logging and webserving effort are inside the docker

One proposal is to use "App RUnner" to deploy the container to address those issues. Plan B will be to create an external volume for the docker to write the logs.

[x] test and deploy AWS app runner
[x] swap instances, otherwise activate plan B: create an external volume and redeploy on EC2

oliche commented 9 months ago

I have deployed on https://alyxapp.internationalbrainlab.org, currently testing with my local one instance. Will update over Xmas to allow for DNS to propagate fully

oliche commented 9 months ago

Documentation here https://github.com/int-brain-lab/iblalyx/blob/main/deployment/DEPLOY_README.md

int-brain-lab / iblalyx

Alyx production docker fails to respond - change recipe for production Alyx #45