hashlab / zeta Goto Github PK

Zeta was created in another time, when we had different needs (and had a different infrastructure), so there's some possible legacy in there.
I'll check with @chrisenytc to validate what can be removed.

• Check if image exists
• Check if repository exists
• List repositories

Try to improve some commands to use user-friendly names instead of IDs (specially deploy command).

Since we're deploying a specific commit, we must validate if both repository and commit exists in GitHub before proceeding with the deploy process.

When merging to master CircleCI reports errors, that some configuration file is missing. Find the complete error here

The API urls are hard-coded, but for greater flexibility, it should be on a separate .env file.

Some things must be done when dealing with Rancher:

• List all available projects
• List project's workloads
• Get a workload description
• Perform actions (rollback, pause and resume)
• Deploy to a workload

The command's documentation will be written here.

We have several workloads at Rancher that we need to individually deploy to, and that's a very tedious works to do. So, to make the deployment process better, we should be able to deploy to all those workloads at once.

We're using Hash Pipe token, but this project should have its own token.

Zeta is not waiting for messages to be sent before proceeding with execution, so when deploying itself, all messages are not sent (since Rancher kills the pods to upgrade its image). But since it only impacts messages when deploying Zeta, that's a very low priority.

Maybe due to dependencies updates, Hubot is throwing an error (that does not block execution) when calling the RespondToUser method at respond.js helper.

Since it does not blocks execution and needs investigation to solve, it will stay there for a while until the task is finished.

To be able to access Zeta from anywhere, without the need to use VPN, the best solution will be to host it at Rancher.