Merge pull request #173 from StevenTobin/fix_observability_config_map

Create Metrics CM in correct place  on core cluster

README.md

@@ -141,7 +141,9 @@ Everything should be shown green. If it is not, click the icon of the faulty obj
   * Enable [monitoring for user-defined projects](https://access.redhat.com/documentation/en-us/openshift_container_platform/4.13/html/monitoring/enabling-monitoring-for-user-defined-projects) in OpenShift clusters
     * `oc -n openshift-monitoring edit configmap cluster-monitoring-config`
     * Set variable `enableUserWorkload` to `true`
+* Core/Hub cluster
   * Edit contents of [thanos-secret](acm/odh-core/acm-observability/secrets/thanos.yaml) file.
+  * Install the ACM observability stack by running `make install`
 
 ## Contributing
 

acm/odh-core/acm-observability/files/metrics_list.yaml

@@ -0,0 +1,24 @@
+names:
+  # MLServer metrics
+  - parallel_request_queue_sum
+  - parallel_request_queue_bucket
+  - parallel_request_queue_count
+  - rest_server_request_duration_seconds_sum
+  - rest_server_request_duration_seconds_bucket
+  - rest_server_request_duration_seconds_count
+  - rest_server_requests_in_progress
+  - rest_server_requests_total
+  - ovms_infer_req_active
+  - ovms_infer_req_queue_size
+  - ovms_inference_time_us_bucket
+  - ovms_inference_time_us_count
+  - ovms_inference_time_us_sum
+  - ovms_request_time_us_bucket
+  - ovms_request_time_us_count
+  - ovms_request_time_us_sum
+  - ovms_requests_fail
+  - ovms_requests_success
+  - ovms_streams
+  - ovms_wait_for_infer_req_time_us_bucket
+  - ovms_wait_for_infer_req_time_us_count
+  - ovms_wait_for_infer_req_time_us_sum

acm/odh-core/acm-observability/kustomization.yaml

@@ -21,3 +21,8 @@ secretGenerator:
   - secrets/thanos.yaml
   options:
     disableNameSuffixHash: true
+- name: observability-metrics-custom-allowlist
+  files:
+  - files/metrics_list.yaml
+  options:
+    disableNameSuffixHash: true