From 69066c90ea08fe7907d0082891a3a7f53c67b213 Mon Sep 17 00:00:00 2001 From: Aleksei Kaplin Date: Thu, 14 Oct 2021 05:15:37 +0200 Subject: [PATCH] TMP fix for ml-pipelines --- .../kcs/install_kubeflow/monitoring.env.yaml | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/automation/infrastructure/terraform/kcs/install_kubeflow/monitoring.env.yaml b/automation/infrastructure/terraform/kcs/install_kubeflow/monitoring.env.yaml index 4f66734..e21d010 100644 --- a/automation/infrastructure/terraform/kcs/install_kubeflow/monitoring.env.yaml +++ b/automation/infrastructure/terraform/kcs/install_kubeflow/monitoring.env.yaml @@ -17,6 +17,12 @@ prometheusOperator: prometheus: prometheusSpec: additionalScrapeConfigs: + # Custom monitoring: monitoring ml-pipeline (the Kubeflow Pipelines API server) + # TODO: remove when fixed: https://github.com/kubeflow/manifests/issues/2011 + - job_name: 'kubeflow-ml-pipeline' + scrape_interval: 15s + static_configs: + - targets: ['ml-pipeline.kubeflow:8888'] # Scan for all services in "kubeflow..." namespaces with prometheus.io annotations (scrape=true, port, address). # Both HTTP and HTTPS @@ -29,30 +35,38 @@ prometheus: kubernetes_sd_configs: - role: service relabel_configs: + # Only "kubeflow..." namespaces - source_labels: [__meta_kubernetes_namespace] action: keep regex: kubeflow(.*) + # Only with prometheus.io/scrape = true - source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scrape] action: keep regex: true + # Use scheme - source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scheme] action: replace target_label: __scheme__ regex: (https?) + # Use path (/metrics) - source_labels: [__meta_kubernetes_service_annotation_prometheus_io_path] action: replace target_label: __metrics_path__ regex: (.+) + # Use address & port - source_labels: [__address__, __meta_kubernetes_service_annotation_prometheus_io_port] action: replace target_label: __address__ regex: ([^:]+)(?::\d+)?;(\d+) replacement: $1:$2 + # Remove prefix from labels - action: labelmap regex: __meta_kubernetes_service_label_(.+) + # Save namespace label - source_labels: [__meta_kubernetes_namespace] action: replace target_label: kubernetes_namespace + # Save service name label - source_labels: [__meta_kubernetes_service_name] action: replace target_label: service_name