Resolve "Implement a retry-on-failure mechanism when check results are not satisfactory"

Closes #77 (closed)

• Add a retries field on Check model
• Update store_check_result() method : update Check instance rather than creating a new one and increment retries field
• Do not run check when related mseed is not present on file-system
• Choose a result backend for celery
• Select all unsatisfactory check results
• Rerun failed checks
• Implement both ways of getting inventory (reading it and downloading it) before removing instrumental response
• log BadRequests
• Make sure that Check is updated for every retry (now, the store_check_result() method is not run)
• Add a daily task to the beat_schedule to rerun problematic checks
• Refactoring des tâches get_mseed_for_channel, build_check_list et check_item (supprimer l'argument date ?, )

https://stackoverflow.com/questions/60986977/celery-chaining-parallel-tasks-into-chord#answer-60988733