Currently, the Helm chart supports a global log level setting under the logging key. We should extend the logging configuration to include subsystem-specific log levels to offer better control and customization. This could be structured as follows:
logging:
logLevel:
subsystems:
# Seconds to persist log level overrides for individual subsystems before Redpanda reverts to the default set in `logging.logLevel`.
# If 0, persist until shutdown. Default is 300 seconds.
# https://docs.redpanda.com/current/reference/rpk/rpk-redpanda/rpk-redpanda-admin-config-log-level-set/#flags
expiration:
logLevels:
s3:
rpc:
# Additional subsystems and their desired log levels
This split approach complicates the configuration process and obscures the clarity and accessibility of logging settings. Consolidating all logging configurations under a single logging section would streamline the UX.
What would you like to be added?
Currently, the Helm chart supports a global log level setting under the
logging
key. We should extend the logging configuration to include subsystem-specific log levels to offer better control and customization. This could be structured as follows:All loggers can be found using:
Related issue: https://github.com/redpanda-data/redpanda/issues/17711
Why is this needed?
Configuring logging is currently split between the
logging
stanza andstatefulset.additionalRedpandaCmdFlags
:This split approach complicates the configuration process and obscures the clarity and accessibility of logging settings. Consolidating all logging configurations under a single logging section would streamline the UX.
JIRA Link: K8S-144