Parameter tuning
Apart from the connection-related parameters, there are a handful of parameters that you can tune to improve the performance your Amazon Aurora MySQL-Compatible Edition cluster when faced with hypergrowth. When changing any database parameter, it's important to use the following best practices:
-
Change one parameter at time so that you can measure the impact of the change.
-
Some parameters might require a warmup period to manifest their effect after they are changed. Consider this when you observe and measure the performance of your Aurora MySQL-Compatible cluster.
-
Avoid incorrect parameter values, which could prevent the database instance from starting-up.
These parameters include:
-
sync_binlog -
table_open_cache -
innodb_sync_array_size -
innodb_flush_log_at_trx_commit -
autocommit -
tmp_table_size -
max_heap_table_size
For guidelines on configuring these parameters, see Aurora MySQL configuration parameters, Recommendations for MySQL features in Aurora MySQL, and Temporary table behavior in Aurora MySQL in the AWS documentation.
