unidata / asanabot Goto Github PK

Would be really nice to find some way to inject e-support in here rather than mail forwards direct to Asana.

It's a good workflow to be able to "check off" tasks corresponding to issues/tasks if there is nothing to be done (i.e. waiting on responses/action by others). Need to ensure that when a PR is updated with new code that the existing task is re-opened and appears back in the assignee's list. No idea right now if this happens or not.

Right now because all tasks are created under my account, I end up following everything and getting notified about it. There needs to be code to optionally unfollow created tasks, controlled by our json configuration.

Make GitHub handling more configurable:

• Column names for boards
• Repo mapping
• Does the repo care about milestones
• Default task assignment for issues with no assignee

• Should be one core asana task creation loop
• GitHub- and StackOverflow-specifics should be in classes representing those items that provide appropriate methods to get the information needed
• Would be really nice to have that Asana task loop exist as a single lambda feeding off the SNS pipe, which is powered by StackOverflow and Github specific lambdas. Not strictly necessary since Stack Overflow is just running in a cron job

In addition to adding a task to the main repo project, we should also look at creating a project for the milestone assigned to an issue/pr. Could take the team from the team for the main project. Milestone project could be 'Repo/Main project + Milestone name'

When a PR is made, the task of reviewing it shows up in the freezer. It should be in the Review column.

S3 isn't great for everything we're doing here. It seems to be introducing latency for the GitHub hook. Also, we're doing a lot of puts to cache update times for StackOverflow.

• Move GitHub and Asana client info to SSM? Not sure if we can set this up with Cloud Formation, or even if we should
• Separate StackOverflow update times from config json
• Put update times in DynamoDB?
• Should json config go somewhere other than S3?