integrations/jetstream/documentation.yaml (39 lines of code) (raw):
exporter_type: included
app_name_short: JetStream
app_name: {{app_name_short}}
app_site_name: {{app_name_short}}
app_site_url: https://github.com/AI-Hypercomputer/JetStream
exporter_name: {{app_name_short}}
exporter_repo_url: https://github.com/AI-Hypercomputer/JetStream/blob/main/docs/observability-prometheus-metrics-in-jetstream-server.md
gke_setup_url: /kubernetes-engine/docs/tutorials/serve-llm-tpu-jetstream-pytorch
additional_prereq_info: |
To collect Prometheus-format metrics from {{app_name_short}}, you must first [build and upload a {{app_name_short}} PyTorch server image](https://github.com/AI-Hypercomputer/jetstream-pytorch/tree/main/docker/jetstream-pytorch-server).
additional_install_info: |
To verify that {{exporter_name}} is emitting metrics on the expected endpoints, do the following:
1. Set up port forwarding by using the following command:
<pre class="devsite-click-to-copy">
kubectl -n {{namespace_name}} port-forward {{pod_name}} 9090
</pre>
2. Access the endpoint `localhost:9090/metrics` by using the browser
or the `curl` utility in another terminal session.
podmonitoring_config: |
apiVersion: monitoring.googleapis.com/v1
kind: PodMonitoring
metadata:
name: jetstream
labels:
app.kubernetes.io/name: jetstream
app.kubernetes.io/part-of: google-cloud-managed-prometheus
spec:
endpoints:
- port: 9090
scheme: http
interval: 30s
path: /metrics
selector:
matchLabels:
app: jetstream-pytorch-server
sample_promql_query: up{job="jetstream", cluster="{{cluster_name}}", namespace="{{namespace_name}}"}
dashboard_available: true
multiple_dashboards: false
dashboard_display_name: {{app_name}} Prometheus Overview