Amazon RDS Performance Tuning Settings

gistfile1.sh

rds-modify-db-parameter-group {param-group-name} \
--parameters="name=character_set_server, value=utf8, method=pending-reboot" \
--parameters="name=collation_server, value=utf8_general_ci, method=pending-reboot" \
--parameters="name=tmp_table_size, value={DBInstanceClassMemory/16}, method=pending-reboot" \
--parameters="name=max_heap_table_size, value={DBInstanceClassMemory/16}, method=pending-reboot" \
--parameters="name=query_cache_type, value=1, method=pending-reboot" \
--parameters="name=query_cache_size, value=131072, method=pending-reboot" \
--parameters="name=table_open_cache, value=2500, method=pending-reboot" \
--parameters="name=join_buffer_size, value={DBInstanceClassMemory/64}, method=pending-reboot" \
--parameters="name=thread_cache_size, value={DBInstanceClassMemory/12582880}, method=pending-reboot" \
--parameters="name=sort_buffer_size, value={DBInstanceClassMemory/64}, method=pending-reboot" \
--parameters="name=innodb_flush_log_at_trx_commit, value=0, method=pending-reboot" \
--parameters="name=innodb_additional_mem_pool_size, value={DBInstanceClassMemory/64}, method=pending-reboot" \
--parameters="name=innodb_max_dirty_pages_pct, value=90, method=pending-reboot" \
--parameters="name=max_allowed_packet, value=33554432, method=pending-reboot" \
--parameters="name=tx_isolation, value=READ-COMMITTED, method=pending-reboot"

I would also increase innodb_read/write_io_threads to 16.

And would mention that innodb_flush_log_at_trx_commit at 0 decreases data durability in case of power loss, that should be taken very seriously.

Is the parameter query_cache_size specified in bytes or in kilobytes? A google search have reported mixed results.

If it indeed is in bytes, then the value is to low.