AWS DMS performance degradation until stopped and resumed
Hello, We use DMS to replicate data from Sql Server to Aurora Postgresql. One scenario we've found recently is that the inserts and updates on sql server are slow to get applied to postgres. The odd thing is that we can stop the task, resume it, and it will catch up immediately. It doesn't happen in particularly busy times, its often coped very well with large loads, however, in quieter periods it can seem to replicate very few inserts/updates/etc over the course of a few hours, fall further behind, but essentially stopping the task and resuming it seems to fix it. Appreciate anyone taking the time to read and think about this question, any additional information I will try to provide. Thanks
I recommend starting with reviewing the task latency, specifically where the latency is. This will help you narrow down what you should be looking at.
If you see source latency on an AWS DMS task, it can be caused by one or more of the following:
- The source database has limited resources.
- The AWS DMS replication instance has limited resources.
- The network speed between the source database and the AWS DMS replication instance is slow.
- AWS DMS reads new changes from the transaction logs of the source database during ongoing replication.
- AWS DMS task settings are inadequate or large objects (LOBs) are being migrated.
High CDCLatencyTarget is most likely the result of the delay in capturing the change events from the source. If the CDCLatencySource isn't high, but the CDCLatencyTarget is high, the latency could be caused by the following:
- There are no primary keys or indexes in the target
- There are resource bottlenecks in the target
- There are resource bottlenecks in the replication instance
- There is a network issue between the replication instance and the target
Guidance on troubleshooting can be found at How do I troubleshoot high source latency on an AWS DMS task? and How can I troubleshoot high target latency on an AWS DMS task?
If you're able to identify where the latency starts (source/target), reviewing the Performance Insights dashboard for the instance might also help you find any bottlenecks or review counter metrics that may help explain what could be contributing to the slowdown.
DMS from Aurora to RedshiftAccepted Answerasked 5 years ago
Questions on Data lake using DMSAccepted Answerasked 3 years ago
Does AWS DMS support ARRAY data type for RDS for PostgreSQL on EC2 to Aurora PostgreSQL migration?Accepted Answerasked 2 years ago
DMS to migrate SQL Server from RDS to EC2asked 2 years ago
DMS CDC for SQL Server - ReplicationAccepted AnswerEXPERTasked 2 years ago
AWS DMS performance degradation until stopped and resumedasked 2 months ago
What happens to a DMS job when a SQL Server AG fails over?Accepted Answerasked 2 years ago
Data Base Migration from On Prem My SQL Server to AWS RDS using DMSasked a month ago
Phantom Update on Aurora Postgres from AWS DMS Replication from SQL Server Sourceasked 3 months ago
Very Large ASYNC_NETWORK_IO on SQL Server when migrating data from SQL Server EC2 to Aurora Postgres using DMSasked 5 months ago