We started using your application and there is a huge transition process where some questions/issues appeared. At this point we are unable to tell if there is or will be a solution.
1, If the Redash container throw an connection error (eg. Cannot connect to database) then the query couldn’t run on a daily basis, even after the Redash can connect to the database. It only can run on the next day when it should run.
We received the below answer for the refresh schedules:
“Refresh schedules are evaluated ever thirty seconds. This means that every thirty seconds, Redash checks whether any scheduled queries need to be executed. It does this by subtracting the most recent execution timestamp from the current time. If the difference is greater than or equal to the scheduled interval then an execution is enqueued. If the difference is less than the scheduled interval then the query is not executed.
This means that scheduling a query to run every thirty minutes means it will progressively execute later and later in the day. Because the query itself takes some period of time to execute (could be milliseconds or minutes). Query run-time isn’t deterministic so this can’t be avoided.
The guarantee a scheduled query makes is that the most recently cached result is never more than N + execution time minutes old, where N is the interval you pick.”
Based on the answers, it should run.
2, The other problem is belonging to the DST(Daylight saving time). For example: We have set the query to 8 o’clock at every day.
The time change came in end of October, the Redash set automatically to 7 o’clock at every day.
Is there a global solution for that?