When write-ahead log files fill up, a checkpoint is performed to force all dirty buffers to disk so the log file can be recycled Checkpoints are also performed automatically at certain intervals, usually every 5 minutes. If there is a lot of database write activity, the write-ahead log segments can fill too quickly, causing excessive slowness as all dirty disk buffers are flushed to disk.