BPT execution logging purging/cleaning done internally by the platform


In some large customers, and projects that work massively with BPT, we are facing sometimes performance issues related with lots of data stored in the instance execution tables and that may impact the platform execution in certain cases. For example some pages on service center tend to start giving timeouts when looking into process data or by going to the monitor Health screen.

This data can be good for short time analysis, but having process execution data for more than one month  does not provide great benefits to the customers/users.

Of course we can use the DBCleaner application and the API. But in most cases when we use this tool, the logging tables are already too full, and we sometimes we face issues removing any old data.

My ideia was to implement something similar to what already exists with the Error/General Log concept (with rotating log tables, where terminated processes are purged after a while). or some sort of purging mechanism that maintains the system clean and stable. 

This can be enhanced by a configuration that could let customers to enable disable this feature if they wanted to maintain this logs. 

Created on 26 Feb
Comments (1)

Changed the category to Backend