This page covers solutions for aggregated and centralized logging collection.
Not yet supported.
Logs are aggregated automatically at the cluster and pushed to the CERN IT logging infrastructure (timber).
A few parameters can be controlled via labels in the cluster definition (and cluster template).
The most relevant ones are:
- logging_producer: The 'producer' of the monitoring data, similar to what you usually provide when configuring CERN monitoring tools. You might already have one, but if not ask the CERN IT monitoring team
- logging_include_internal: True if the internal cluster service logs (kube-system namespace) should also be sent to the configured endpoint (default is False)
To create a cluster with a custom logging_producer, pass the value of the 'producer' you want to use.
As an example:
$ openstack coe cluster create <name> --cluster-template <cluster-template> \ --merge-labels \ --labels logging_producer=MYPRODUCER