Missing Environment Variables when running _deploy.sh in deploy/env-prod-v3

[Skeeve4711]

Hello,

I am having trouble to set klicker v3 up on a local instance. When I run the command
doppler run --config prd -- helmfile build -f ./helmfile.yaml in deploy/env-prod-v3
I get the following error message
in ./helmfile.yaml: error during helmfile.yaml.part.0 parsing: template: stringTemplate:15:23: executing "stringTemplate" at <requiredEnv "APP_SECRET">: error calling requiredEnv: required env var `APP_SECRET` is not set
I can set that variable but then I'll also need to set a lot of other variables and with some of them I am unsure. Is there an example environment somewhere so I know how it's supposed to look like?

Thank you in advance!

[rschlaefli]

Hi @Skeeve4711,

We have just released the v3 for public usage at our home university, so the documentation and examples for deployment are not yet available, and the deployment is still customized to our environment using Microsoft Azure. We will be adding more information and examples regarding self-hosting of v3 after we have been in production for a while.

It would be very helpful for us if you could shortly describe your hosting use case (where (Cloud, Kubernetes, ...) and for what use case and audience you would like to host v3). We would like to account for your needs as well once we start improving on the self-hosting experience.

Furthermore, some core functionalities (i.e., processing responses in a live quiz, uploading images) are currently closely tied to Azure functionalities (Azure Functions and Blob Storage), and will need to be generalized to be usable in another environment (e.g., using only Docker Compose or Kubernetes).

KlickerUZH v2 (dev branch) is a much less complex application, and we have more deployment examples for that version, but the migration scripts from v2 to v3 are currently only available as Azure Functions. So it might be best to use v3 from the start unless you are using Azure and can set up the migration for your users as well.

[Skeeve4711]

Hi @rschlaefli,

Thank you for your reply! We will just wait for now then. V2 isn't really worth the hazzle because of migrating afterwards.

Is there maybe a time window on when you approximately plan to improve the self hosting?

We want to host v3 on a virtual machine here and would prefer to use Docker compose if possible and we want to use it in lectures here for live quizzes during it.

[rschlaefli]

Hi @Skeeve4711,

Thanks for the additional context. I think we will first need to worry about finalizing everything in terms of documentation (also explaining the new architecture), and then we can create clean examples for, e.g., Docker Compose based hosting.

We will need to find alternatives for components like Azure Functions, but with a local VM and especially in smaller use cases that could be handled directly by the backend. We only use serverless because our instance is public and large.

In terms of timeline, we can probably start evaluating this again in late October, after finishing all documentation-related tasks and ensuring the platform is running smoothly in the first few weeks of our semester. Does that work for your use case?

I will leave this open and come back to you once we start thinking about this topic.

Thanks for the interest in our project!

Best,
Roland

[Skeeve4711]

HI @rschlaefli,

that works for us. We'll just wait for now then.

[rschlaefli]

Hi @Skeeve4711,

Quick update: we have started documenting the local development set-up at https://github.com/uzh-bf/klicker-uzh/wiki/Local-Development. There is also an extended docker-compose.yml example available at https://github.com/uzh-bf/klicker-uzh/blob/v3/docker-compose.yml. While this does not yet completely satisfy your request, it might be a start if you want to take a look at the app and its dependencies/set-up.

We will continue working on the documentation in the coming weeks and add more examples related to deployment. Also, we will still need to find alternatives for some Azure-only services (e.g., the Azure functions, Azure service bus, and Azure blob storage) before you can run a productive workload in a VM.

Hope that gives an initial indication and feel free to ask if anything is unclear, so that we can improve the documentation with your feedback.

Best,
Roland

[Skeeve4711]

Hi @rschlaefli,

that's already a good start. Thank you! We can look into it already to get to know a little bit about the architecture but we will still wait before deploying it until you hopefullly find alternatives for the Azure-only services.

Best regards,
Pascal