is there a way to Clean up a table with 50Million records without impacting performance

[P Seth]

Hello all , 
Recently i came across the incident activitylog table which has nearly 60M records and since this data is not required and its growing , we planned to keep only 3year data and everything older than 3year has to be deleted and table should be maintained with 3year data only . 
for maintaining we can use table cleaner , but i am confused about first-time delete will table cleaner will be able to do it without impacting the performance ?? 
what should be the best practise to be implemented here

[GlideFather · Accepted solution]

you can do it via scheduled jobs scheduled on weekends and/or after business hours, and by smaller parts e.g. setLimit(100000);

Also, consider this: Data archiving (ServiceNow documentation).

[GlideFather · Accepted solution]

you can do it via scheduled jobs scheduled on weekends and/or after business hours, and by smaller parts e.g. setLimit(100000);

Also, consider this: Data archiving (ServiceNow documentation).

[JC Moller · Accepted solution]

If it is a regular table, then just add a Table Cleaner job to handle the deletion. The Table Cleaner starts once a hour and deletes data for 20 minutes. Never had any issues even with larger tables. You can off-course set the filter condition to start from a time period longer back in time to make the deletable data amount smaller, eg. sys_created_on relative before xx days. And then update the condition as the deletes proceed. I usually do a glideaggregate/trend by year-month analysis if I need to delete a larger amount for the first time, so I know how to build the filter conditions if I wanna play it safe and just delete 100K at each round.

BR, Jan