How long to keep MySQL binary logs for CDC

A CDC pipeline reads MySQL's binary log from a saved position. If it stops (an outage, a network change, a paused pipeline) and MySQL deletes the logs it hasn't read yet, it can't continue: the tables have to be copied again.

What to set

Keep at least 3 days, ideally 7, so a problem that starts on a Friday evening can be fixed on Monday. Binary logs take disk space roughly proportional to your write volume, so check free storage after raising it.

By host

• Self-hosted MySQL 8: SET PERSIST binlog_expire_logs_seconds = 604800; (7 days; MySQL 8's default is 30 days).
• Amazon RDS and Aurora: CALL mysql.rds_set_configuration('binlog retention hours', 168); and check with CALL mysql.rds_show_configuration;. Without it, RDS can remove binary logs soon after they're written.
• Google Cloud SQL: binary logs are kept for the point-in-time recovery window; set the transaction log retention in days on the instance.
• Azure Database for MySQL: set the binlog_expire_logs_seconds server parameter.

How Quayen helps

The connection test checks the retention and warns if it's too short. If logs were already purged when a pipeline resumes, Quayen says so plainly and lets you re-copy the affected tables instead of silently skipping changes.