Description
Disabling log_checkpoints reduces diagnostic information about checkpoint and restartpoint duration and buffer writes. It does not stop the checkpoints themselves.
Potential impact
Tracing spikes in disk I/O or performance degradation can become harder. Missing operational evidence can delay incident investigation and database tuning.
Remediation
Set log_checkpoints to on on current Flexible Server. Verify actual log collection and retention, and restrict access to the logs. Configure logs for other purposes, such as query auditing, separately.
Examples
The initial excerpt is historical configuration for the retired Single Server service. In the revised excerpt, MyDBServer1 refers to a separately prepared Flexible Server; its parent declaration and data migration are omitted.
Before
resource MyDBServer1_log_checkpoints 'Microsoft.DBforPostgreSQL/servers/configurations@2017-12-01' = {
parent: MyDBServer1
name: 'log_checkpoints'
properties: {
value: 'off'
}
}
Checkpoint diagnostic logging is disabled.
After
resource MyDBServer1_log_checkpoints 'Microsoft.DBforPostgreSQL/flexibleServers/configurations@2024-08-01' = {
parent: MyDBServer1
name: 'log_checkpoints'
properties: {
source: 'user-override'
value: 'on'
}
}
Checkpoint logging is enabled. After applying the setting, verify that collected logs include duration and write information.