Details
-
Feedback
-
Resolution: Won't Do
-
Major - P3
-
None
-
None
-
None
-
None
Description
Description
During the push-based live migration of the client([Ignite sales|https://mongodb.lightning.force.com/lightning/r/Account/001A000001B7WXAIA3/view),] the Live migration(Push) method slowed down the application queries and started causing time outs in the application side.
This was happening because the client was using gp2 EBS volume with 350GB storage and once the migration was started the read IOPS peaked from 500/sec to 2.5k/sec however after the burst window it came down to 500 IOPS again. Once that happened the entire production cluster slowed down and the disk latency and disk queue depth increased significantly.
Changes
The ask here is to include this IOPS consideration in the considerations section of the live migration(ref - https://www.mongodb.com/docs/atlas/import/migrate-from-com-rs/#considerations) as it is a pretty critical check while running the production migration and can cause significant downtime.
Screenshot -
* *