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
We are running docker-postfix in K8s using the provided HelmChart.
Since we have activated the prometheus metrics, an additional postfix log file is generated for the prometheus exporter.
In our productive environment we had the problem that Postfix did not write anything to the new log after the log update and therefore not all metrics were available anymore.
We have now added the "copytruncate" parameter to the logrotate configuration in values.yaml.
Since then there have been no more problems.
IMHO it makes sense to add this parameter to the values.yaml.
The text was updated successfully, but these errors were encountered:
We are running docker-postfix in K8s using the provided HelmChart.
Since we have activated the prometheus metrics, an additional postfix log file is generated for the prometheus exporter.
In our productive environment we had the problem that Postfix did not write anything to the new log after the log update and therefore not all metrics were available anymore.
We have now added the "copytruncate" parameter to the logrotate configuration in values.yaml.
Since then there have been no more problems.
IMHO it makes sense to add this parameter to the values.yaml.
The text was updated successfully, but these errors were encountered: