Last Activity: 4 months ago
Smarter Exception Logging
Your idea has been submitted and is awaiting moderation.
A simple and non-critical error in a frequently running task can easily flood the exception log. This has two detrimental effects:
1. It is very hard for end users to determine errors in the system when this happens as the frequent error obscures any other errors.
2. The error log size can bring down the system.
I suggest we decouple the log itself from the moment the error occurred and log the latter in a separate table. The exception logic would need to check for the specific error first. This could be done by storing a hash of the error and using that as unique key in the error log table. And then recording hash and time in the new table.
Then, in the error log UI give the user 2 view options:
1. Show errors sequentially as they occurred (i.e. as it is now)
2. Show the most occurring errors in descending order