massover / killington Goto Github PK

Instead we can get the information we need from the count of enetered users.

Also have to update the ActiveLotterManager and change uses of Lottery.active_objects.all(). Instead uses should query by users that have show subscriptions to that have active lotteries that they have not been entered in.

After signing up, no one knows what to do

1. Pass kwargs via extra params, eg. sentry docs

2. Make it easy to debug directly from sentry, instead of relying on paper trail

3. Move the enter_lottery HTML response to another place(possibly s3 with links for easy access?)

4. Use celery task id as identifier for easy log searching.

Given I am an oauth2 authenticated client
When I make a request to the user endpoint by SES email
Then I should be able to get the user's real email

https://docs.sentry.io/clients/python/transports/#sync
getsentry/raven-python#881

Raven does not send exceptions/logs with async transport mode when CELERY_WORKER_MAX_TASKS_PER_CHILD = 1. Sync HttpTransport fixes it and works fine, but it's
obviously at the cost of being sync. This should be removed and set back to the default if it gets fixed:

# settings/base.py

from raven.transport import HTTPTransport
RAVEN_CONFIG = {
    'transport': HTTPTransport
}

• Upon start send an email
• Upon start add task to queue for failure email at end_time + 10 minutes(to figure out exact time later)
• Add a success field to the flood
• Add api for Flood to post success = True
• Add api for getting flood_id, probably from SESViewSet
• Add failure email task which queries a flood, checks if it's a success, and emails if it was not a success.

... or a human can do it

run_shows_spider must be run before lottery processing tasks can run, so let's run it at the first minute of every hour

run_shows_spider at minute=1, hour='*'

Confirm how long the task usually takes using logs and then kick off the lottery processing a little after.

If the lottery is not active anymore because a task is taking too long, exit.