Logs in Managed Service for ClickHouse®

This article provides information on concepts that will help you understand and better manage logs for your DoubleCloud cluster:

Log entries

Log entry records status or describes certain events in your cluster. These messages tell you that certain events occur in the cluster or show something is wrong and needs attention.

Log entry structure

A log entry contains the complete information on a logged event and consists of the following segments:

Date and time stamp in YYYY/MM/DD, HH:MM:SS format.
Log level indicator.
Entry text.

Log levels

Log level is a piece of status information that tells you how important a log entry is.

Your DoubleCloud cluster supports the following log levels:

Level	Description
`TRACE`	Detailed diagnostics information
`DEBUG`	Debugging information
`INFO`	Diagnostics information for statistics
`WARNING`	Warning about a non-critical malfunction. Please investigate further
`ERROR`	Error report
`FATAL`	Possible system failure, cluster shutdown

Default values for system log tables

DoubleCloud Managed Service for ClickHouse® allows you to manage system.*_log tables:

Enable and disable the tables (except system.query_log and system.part_log).
Change the disk space allowed for the table (in bytes) - retention_size.
Change the table partition's retention time (in seconds) - retention_time.

This parameter is used after conversion into days by the following formula:
```
(ceil (table_retention_time / (60 * 60 * 24)))
```
The resulting number determines the limit of partitions for the table. Each table has one partition per day.

Log tables cleanup schedule

Log tables are cleaned up once every 24 hours, after 0:00 according to the timezone of the cluster's region. This allows the peak usage to exceed the set limits.

system.query_log

You can't turn off this table

You can't apply system.query_log_enabled: false, only change the size and time.

`query_log_retention_size`	1073741824
`query_log_retention_time`	2592000

system.query_thread_log

`query_thread_log_enabled`	true
`query_thread_log_retention_size`	536870912
`query_thread_log_retention_time`	2592000

system.query_views_log

`query_views_log_enabled`	false
`query_views_log_retention_size`	0
`query_views_log_retention_time`	2592000

system.part_log

You can't turn off this table

You can't apply system.part_log_enabled: false, only change the size and time.

`part_log_retention_size`	536870912
`part_log_retention_time`	2592000

system.metric_log

`metric_log_enabled`	true
`metric_log_retention_size`	536870912
`metric_log_retention_time`	2592000

system.asynchronous_metric_log:

`asynchronous_metric_log_enabled`	false
`asynchronous_metric_log_retention_size`	0
`asynchronous_metric_log_retention_time`	2592000

system.text_log

`text_log_enabled`	false
`text_log_retention_size`	536870912
`text_log_retention_time`	2592000
`text_log_level`	trace

system.trace_log

`trace_log_enabled`	false
`trace_log_retention_size`	536870912
`trace_log_retention_time`	2592000

system.opentelemetry_span_log

`opentelemetry_span_log_enabled`	false
`opentelemetry_span_log_retention_size`	0
`opentelemetry_span_log_retention_time`	2592000

system.session_log

`session_log_enabled`	false
`session_log_retention_size`	0
`session_log_retention_time`	2592000

system.zookeeper_log

`zookeeper_log_enabled`	false
`zookeeper_log_retention_size`	0
`zookeeper_log_retention_time`	2592000

system.asynchronous_insert_log

`asynchronous_insert_log_enabled`	false
`asynchronous_insert_log_retention_size`	0
`asynchronous_insert_log_retention_time`	2592000