As we prepare for a 1.0 release, we're taking taking to opportunity to remove various inconsistencies and warts in Prometheus.
There are two such changes in this release. The first is that
target_groups has been renamed to
static_configs, backwards compatibility will be kept in this release. This is consistent with all other service discovery configs.
names has been renamed to
files. This is not backwards compatible. This is to be consistent with
Here's a config file showing examples of the new names:
scrape_configs: - job_name: 'node' static_configs: # This used to be "target_groups" - targets: - localhost:9090 - job_name: 'dummy' file_sd_configs: - files: # This used to be "names" - targets.json
There's also some more minor changes to the configuration. It's now possible to configure the server name used with TLS. Kubernetes now takes the node port from the kubelet rather than configuration, and provides additional information for relabelling.
Finally in preparation for the HA Alertmanager which is currently under code review, it's now possible to specify a comma separated list of URLs to
-alertmanager.url. This will send alerts to all the provided alertmanagers.
There were as usual a variety of of fixes and minor improvements. For the full list of changes, see the release notes.