Description
log_checkpoints records checkpoints and restartpoints, including statistics such as buffers written and time spent writing. Without these logs, investigating I/O spikes or unusual checkpoint frequency is harder. Disabling this logging does not stop checkpoints themselves.
Potential impact
- It may be harder to relate disk-write load to response delays.
- Incident investigation and performance tuning may lack useful records.
Remediation
- Set
log_checkpointstoonin the PostgreSQL service in use. In Azure, use that service’s server-parameter settings. - Review the logs alongside performance metrics and verify collection and retention.
Examples
The first example targets Single Server, which retired on March 28, 2025. The second applies to an existing Flexible Server and uses a module available in azure.azcollection 3.18.0 and later.
Before
yaml
- name: PostgreSQL 설정 변경
azure.azcollection.azure_rm_postgresqlconfiguration:
resource_group: myResourceGroup
server_name: myServer
name: log_checkpoints
value: "off"
Checkpoint logging is disabled. Checkpoints still occur, but the diagnostic records provided by this setting are absent.
After
yaml
- name: PostgreSQL 설정 변경
azure.azcollection.azure_rm_postgresqlflexibleconfiguration:
resource_group: myResourceGroup
server_name: myServer
name: log_checkpoints
value: "on"
Checkpoint logging is enabled. Use the actual server and authentication configuration, and verify that logs are collected.