Closed gabrielwol closed 2 months ago
However I'm assuming changing the ecocounter_pull
time to 10am as part of #983 will solve most of the issue so we should evaluate if it's still necessary after that change.
Would it make sense to have all locations as separate tasks with retries?
Definitely open to that!
Here's an example of a sensor that reported 5 days of backlogged data at once.
I successfully tried out dynamically mapped tasks, however need to think more about:
Chris identified another example which can be used for testing; site_id = 300028433
(Yonge St, South of Davisville Ave) transmitted recent data in batches (07/08 and 07/14):
According to Ecocounter, data can sometimes arrive late due to poor cellular connection. We should add a process to retry sensors missing data on previous days (up to a week?).