Here's our current configuration for openstack service discovery for node exporter:
scrape_configs:
- job_name: 'openstack_service_discovery'
openstack_sd_configs:
- identity_endpoint: '
http://openstack.somewhere.org:5000/v3'
role: instance
region: RegionOne
username: <username>
password: <password>
domain_name: default
project_name: project
relabel_configs:
- source_labels: [__meta_openstack_private_ip]
regex: ^172\.18\..*
action: keep
- source_labels: [__meta_openstack_tag_node_exporter]
regex: ^true
action: keep
- source_labels: [__address__]
regex: (.*):80
target_label: __address__
replacement: $1:9100
- source_labels: [__meta_openstack_instance_name]
target_label: openstack_instance_name
We'd like to add a relabel configuration to detect cadvisor exporter targets that may or may not exist on the same nodes on port 8080. Is it possible to do this without having to set up a new job and thus double our openstack api calls? Or triple them in the future if we add another exporter, etc.? Is it somehow possible to do multiple relabel configurations that re-use the data from the one job? What would the syntax look like? We basically want to add (differences highlighted):
- source_labels: [__meta_openstack_private_ip]
regex: ^172\.18\..*
action: keep
- source_labels: [__meta_openstack_tag_
cadvisor]
regex: ^true
action: keep
- source_labels: [__address__]
regex: (.*):80
target_label: __address__
replacement: $1:
8080 - source_labels: [__meta_openstack_instance_name]
target_label: openstack_instance_name