Machine Config Daemon metrics

The Machine Config Daemon is a part of the Machine Config Operator. It runs on every node in the cluster. The Machine Config Daemon manages configuration changes and updates on each of the nodes.

Machine Config Daemon metrics

Beginning with OKD 4.3, the Machine Config Daemon provides a set of metrics. These metrics can be accessed using the Prometheus Cluster Monitoring stack.

The following table describes this set of metrics.

Metrics marked with in the Name* and Description columns represent serious errors that might cause performance problems. Such problems might prevent updates and upgrades from proceeding.

While some entries contain commands for getting specific logs, the most comprehensive set of logs is available using the oc adm must-gather command.

Table 1. MCO metrics
NameFormatDescriptionNotes

mcd_host_os_and_version

[]string{“os”, “version”}

Shows the OS that MCD is running on, such as Fedora.

ssh_accessed

counter

Shows the number of successful SSH authentications into the node.

The non-zero value shows that someone might have made manual changes to the node. Such changes might cause irreconcilable errors due to the differences between the state on the disk and the state defined in the machine configuration.

mcd_drain

{“drain_time”, “err”}

Logs errors received during failed drain.

While drains might need multiple tries to succeed, terminal failed drains prevent updates from proceeding. The drain_time metric, which shows how much time the drain took, might help with troubleshooting.

For further investigation, see the logs by running:

$ oc logs -f -n openshift-machine-config-operator machine-config-daemon-<hash> -c machine-config-daemon

mcd_pivot_err

[]string{“pivot_target”, “err”}

Logs errors encountered during pivot.

Pivot errors might prevent OS upgrades from proceeding.

For further investigation, run this command to access the node and see all its logs:

$ oc debug node/<node> — chroot /host journalctl -u pivot.service

Alternatively, you can run this command to only see the logs from the machine-config-daemon container:

$ oc logs -f -n openshift-machine-config-operator machine-config-daemon-<hash> -c machine-config-daemon

mcd_state

[]string{“state”, “reason”}

State of Machine Config Daemon for the indicated node. Possible states are “Done”, “Working”, and “Degraded”. In case of “Degraded”, the reason is included.

For further investigation, see the logs by running:

$ oc logs -f -n openshift-machine-config-operator machine-config-daemon-<hash> -c machine-config-daemon

mcd_kubelet_state

[]string{“err”}

Logs kubelet health failures.

This is expected to be empty, with failure count of 0. If failure count exceeds 2, the error indicating threshold is exceeded. This indicates a possible issue with the health of the kubelet.

For further investigation, run this command to access the node and see all its logs:

$ oc debug node/<node> — chroot /host journalctl -u kubelet

mcd_reboot_err

[]string{“message”, “err”}

Logs the failed reboots and the corresponding errors.

This is expected to be empty, which indicates a successful reboot.

For further investigation, see the logs by running:

$ oc logs -f -n openshift-machine-config-operator machine-config-daemon-<hash> -c machine-config-daemon

mcd_update_state

[]string{“config”, “err”}

Logs success or failure of configuration updates and the corresponding errors.

The expected value is rendered-master/rendered-worker-XXXX. If the update fails, an error is present.

For further investigation, see the logs by running:

$ oc logs -f -n openshift-machine-config-operator machine-config-daemon-<hash> -c machine-config-daemon

Additional resources