Both the API service and the collector service expose all metrics generated by the standard Go client for Prometheus,
e.g. go_memstats_alloc_bytes for memory usage or go_goroutines for goroutine count.
The API service exposes the standard HTTP metrics generated by the standard Go client for Prometheus:
| Type | Metric | Labels |
|---|---|---|
| Summary | http_request_duration_microseconds |
|
| Summary | http_request_size_bytes |
|
| Counter | http_requests_total |
code, method |
| Summary | http_response_size_bytes |
The collector service exposes the following metrics by default:
| Type | Metric | Labels |
|---|---|---|
| Counter | limes_successful_scrapes |
service, service_name (counts projects) |
| Counter | limes_failed_scrapes |
service, service_name (counts projects) |
| Counter | limes_successful_domain_discoveries |
none |
| Counter | limes_failed_domain_discoveries |
none |
| Counter | limes_successful_project_discoveries |
domain, domain_id |
| Counter | limes_failed_project_discoveries |
domain, domain_id |
| Counter | limes_successful_capacity_scrapes |
capacitor |
| Counter | limes_failed_capacity_scrapes |
capacitor |
| Counter | limes_successful_auditevent_publish |
none |
| Counter | limes_failed_auditevent_publish |
none |
The limes_failed_scrapes metric is particularly useful for assessing the continued operation of backend services
(specifically their API parts). If you can do only one alert on Limes metrics, alert on limes_failed_scrapes.
Alerts on limes_failed_{domain,project}_discoveries are very useful, too, but less important.
For the scraping metrics, the service label contains the type of the backend service in question (as stated in the Keystone
service catalog), and the service_name label contains the product name (in lower case) of the reference implementation
of this service, for instance, service_name="nova" for service="compute".
With the configuration option collector.data_metrics set to true, the collector service will report the following
additional metrics:
| Type | Metric | Labels |
|---|---|---|
| Gauge | limes_cluster_capacity |
service, resource |
| Gauge | limes_domain_quota |
service, resource, domain, domain_id |
| Gauge | limes_project_backendquota |
service, resource, domain, domain_id, project, project_id |
| Gauge | limes_project_quota |
service, resource, domain, domain_id, project, project_id |
| Gauge | limes_project_usage |
service, resource, domain, domain_id, project, project_id |
Plugins may emit their own metrics when enabled. Refer to the documentation of the quota plugins and the capacity plugins for details.