You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description
Our Minio logs bucket is filling up- specifically subdirectories not associated with packet logs.
We have tooling in plugin.txt to configure cleanup microservices based on time and size of packet logs.
We seem to lack similar functionality for limiting logs for text_logs and tool_logs, as well as NOSCOPE logs.
Minio has time-based bucket lifecycle management but these are less powerful and scalable than something based on filesize, which the cleanup microservices can do.
With bucket sizes constrained by volume mounts (like kubernetes PVCs) it is difficult to correlate time-based cleanup policies with volume size, without massively oversizing volumes.
We would like capability similar to the CMD_LOG_CYCLE_TIME, CMD_LOG_CYCLE_SIZE, to apply to all logs cosmos writes to the logs bucket.
Proposed Solution
Allow configuring the built in cleanup microservice to apply to all content in the logs bucket.
This likely needs to live somewhere outside plugin.txt. A global config file? init.sh?
Additional context
We can configure new minio cleanup policies through the cosmos-init container in the short term, and increase the size of our PVCs.
The text was updated successfully, but these errors were encountered:
Description
Our Minio logs bucket is filling up- specifically subdirectories not associated with packet logs.
We have tooling in
plugin.txt
to configure cleanup microservices based on time and size of packet logs.We seem to lack similar functionality for limiting logs for
text_logs
andtool_logs
, as well asNOSCOPE
logs.Minio has time-based bucket lifecycle management but these are less powerful and scalable than something based on filesize, which the cleanup microservices can do.
With bucket sizes constrained by volume mounts (like kubernetes PVCs) it is difficult to correlate time-based cleanup policies with volume size, without massively oversizing volumes.
We would like capability similar to the CMD_LOG_CYCLE_TIME, CMD_LOG_CYCLE_SIZE, to apply to all logs cosmos writes to the logs bucket.
Proposed Solution
Allow configuring the built in cleanup microservice to apply to all content in the logs bucket.
This likely needs to live somewhere outside
plugin.txt
. A global config file?init.sh
?Additional context
We can configure new minio cleanup policies through the cosmos-init container in the short term, and increase the size of our PVCs.
The text was updated successfully, but these errors were encountered: