Most performance issues are linked to database maintenance. Here are four main leads to help you find the cause of slow performance:
Check that the initial Adobe Campaign platform configuration is still valid and if necessary, reassess your customer’s needs in terms of deliverability or database size. We also recommend running a full hardware check (CPU, RAM, IO system).
You can refer to Adobe Campaign Harware Sizing guide for insights.
Inappropriate configuration may affect platform performance. We recommend that you check network configuration, platform deliverability options as well as MTA configuration in the serverConf.xml file.
Database cleanup task
Please make sure the database cleanup task is operational. To do this, view the log files to see if they contain any errors. For more on this, refer to this section.
Make sure database maintenance is correctly scheduled and executed. To do this, please contact your database administrator to learn more about:
For more on this, refer to this section.
If you are using a mid-sourcing configuration, it is essential for databases to be maintained on a regular basis. When analyzing a delivery on the marketing platform, the marketing instance sends information to the mid-sourcing instance. If the process is slowed down, the marketing instance will be impacted.
Managing work tables
Please check the number and size of work tables. When they exceed a certain size, database performance is affected. These tables are created by workflows and deliveries. They remain in the database while workflows and deliveries are active. To limit the size of work tables, you can carry out the following operations:
If the operation takes a long time and frees up a lot of space, this means that in-depth maintenance is necessary (index rebuilding, etc.). For more on this, refer to this section.
Adobe Campaign process monitoring
Depending on Adobe Campaign installation settings, two tools can be used for platform monitoring:
It may become necessary to run a real-time diagnosis to identify the cause of the issue. Start by checking the process and platform log files, then monitor database activity while recreating the issue. Pay particular attention to the following: