feat: label components with /component label, use /name to refer to benchmark pods; add README

k8s/README.md

@@ -0,0 +1,48 @@
+This folder contains the required Kubernetes and Argo Workflow resources required to run experiments in Kubernetes
+both in local (e.g. Minikube, Kind) and remote clusters.
+
+## Prerequisites
+
+### Argo Workflows
+
+Whatever cluster you choose must be running [Argo Workflows](https://argo-workflows.readthedocs.io/).
+
+**Local clusters.** For local clusters, you can follow the instructions in
+the [Argo Workflows Quickstart Guide](https://argo-workflows.readthedocs.io/en/latest/quick-start/) to get Argo
+Workflows running.
+
+For remote clusters, it's best to consult the Argo
+Workflows (Operator Manual)[https://argo-workflows.readthedocs.io/en/latest/installation/].
+
+**Argo CLI Tool.** You will also need to install the
+[Argo CLI tool](https://argo-workflows.readthedocs.io/en/latest/walk-through/argo-cli/) to submit workflows.
+
+**Permissions.** Codex workflows assume that they are running in a namespace called `codex-benchmarks`. We
+have a sample manifest which creates the namespace as well as the proper service account with RBAC
+permissions [here](./argo-workflows/codex-workflows-rbac.yaml). For local clusters, you can apply this manifest
+as it is. For remote clusters, you might need to customize it to your needs.
+
+### Logs
+
+Experiments require logs to be stored for later parsing during analysis. For local clusters, this can be achieved
+by running [Vector](https://vector.dev/) and outputting pods logs to a persistent volume. The manifests for setting the
+persistent volume, as well as vector,
+can be found [here](./local).
+
+### Submitting Workflows
+
+Once everything is set up, workflows can be submitted with:
+
+```bash
+argo submit -n argo ./deluge-benchmark-workflow.yaml
+```
+
+for local clusters, you should add:
+
+```bash
+argo submit -n argo ./deluge-benchmark-workflow.yaml --insecure-skip-verify
+```
+
+To observe progress, you can use the Argo Wokflows UI which can be accessed by port-forwarding the Argo Workflows 
+server.
+

k8s/argo-workflows/deluge-benchmark-workflow.yaml

@@ -177,7 +177,7 @@ spec:
         command: [ "/bin/bash" ]
         source: |
           ./docker/bin/kubectl-wait-job\
-            --selector=app.kubernetes.io/name=deluge-experiment-runner,\
+            --selector=app.kubernetes.io/component=deluge-experiment-runner,\
           app.kubernetes.io/instance=e{{inputs.parameters.runId}},\
           app.kubernetes.io/part-of=g{{inputs.parameters.groupId}}\
             --timeout={{workflow.parameters.maxExperimentDuration}}\

k8s/charts/deluge/templates/_helpers.tpl

@@ -43,6 +43,7 @@ Common and selector labels.
 
 {{- define "deluge-benchmarks.labels" -}}
 helm.sh/chart: {{ include "deluge-benchmarks.chart" . }}
+app.kubernetes.io/name: {{ include "app.name" . }}
 {{ include "deluge-benchmarks.selectorLabels" . }}
 {{- if .Chart.AppVersion }}
 app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}

k8s/charts/deluge/templates/deluge-service.yaml

@@ -4,12 +4,12 @@ metadata:
   name: deluge-nodes-service-{{ include "experiment.fullId" . }}
   namespace: {{ .Release.Namespace }}
   labels:
-    app.kubernetes.io/name: deluge-nodes-service
+    app.kubernetes.io/component: deluge-nodes-service
     {{- include "deluge-benchmarks.labels" . | nindent 4}}
 spec:
   clusterIP: None
   selector:
-    app.kubernetes.io/name: deluge-node
+    app.kubernetes.io/component: deluge-node
     {{- include "deluge-benchmarks.selectorLabels" . | nindent 4}}
   ports:
     - port: 6890

k8s/charts/deluge/templates/deluge-statefulset.yaml

@@ -10,13 +10,13 @@ spec:
   podManagementPolicy: Parallel
   selector:
     matchLabels:
-      app.kubernetes.io/name: deluge-node
+      app.kubernetes.io/component: deluge-node
       {{- include "deluge-benchmarks.selectorLabels" . | nindent 6 }}
 
   template:
     metadata:
       labels:
-        app.kubernetes.io/name: deluge-node
+        app.kubernetes.io/component: deluge-node
         {{- include "deluge-benchmarks.labels" . | nindent 8 }}
 
     spec:

k8s/charts/deluge/templates/testrunner-job.yaml

@@ -4,14 +4,14 @@ metadata:
   name: deluge-experiment-runner-{{ include "experiment.fullId" . }}
   namespace: {{ .Release.Namespace }}
   labels:
-    app.kubernetes.io/name: deluge-experiment-runner
+    app.kubernetes.io/component: deluge-experiment-runner
     {{- include "deluge-benchmarks.labels" . | nindent 4 }}
 
 spec:
   template:
     metadata:
       labels:
-        app.kubernetes.io/name: deluge-experiment-runner
+        app.kubernetes.io/component: deluge-experiment-runner
         {{- include "deluge-benchmarks.labels" . | nindent 8 }}
 
     spec:

k8s/charts/deluge/templates/tracker-deployment.yaml

@@ -4,20 +4,20 @@ metadata:
   name: bittorrent-tracker-{{ include "experiment.fullId" . }}
   namespace: {{ .Release.Namespace }}
   labels:
-    app.kubernetes.io/name: bittorrent-tracker
+    app.kubernetes.io/component: bittorrent-tracker
     {{- include "deluge-benchmarks.labels" . | nindent 4 }}
 
 spec:
   replicas: 1
   selector:
     matchLabels:
-      app.kubernetes.io/name: bittorrent-tracker
+      app.kubernetes.io/component: bittorrent-tracker
       {{- include "deluge-benchmarks.selectorLabels" . | nindent 6 }}
 
   template:
     metadata:
       labels:
-        app.kubernetes.io/name: bittorrent-tracker
+        app.kubernetes.io/component: bittorrent-tracker
         {{- include "deluge-benchmarks.labels" . | nindent 8 }}
 
     spec:

k8s/charts/deluge/templates/tracker-service.yaml

@@ -4,7 +4,7 @@ metadata:
   name: bittorrent-tracker-service-{{ include "experiment.fullId" . }}
   namespace: {{ .Release.Namespace }}
   labels:
-    app.kubernetes.io/name: bittorrent-tracker-service
+    app.kubernetes.io/component: bittorrent-tracker-service
     {{- include "deluge-benchmarks.labels" . | nindent 4 }}
 
 spec:
@@ -14,5 +14,5 @@ spec:
       targetPort: 8000
       protocol: TCP
   selector:
-    app.kubernetes.io/name: bittorrent-tracker
+    app.kubernetes.io/component: bittorrent-tracker
     {{- include "deluge-benchmarks.selectorLabels" . | nindent 4 }}

k8s/local/README.md

@@ -1,2 +0,0 @@
-Configurations which can be applied to a local k8s cluster like Minikube or Kind so that experiment logs can be collected for later parsing. This is a simple version of what one would do in a real cluster: use a log collector and ship everything onto external storage or a service like Logstash.
-

k8s/vector/vector-chart-values.yaml

@@ -5,7 +5,7 @@ service:
 extraVolumes:
   - name: vector-logs
     persistentVolumeClaim:
-      claimName: codex-workflow-logs-pvc
+      claimName: vector-logs-pvc
 
 extraVolumeMounts:
   - name: vector-logs

k8s/vector/vector-pv.yaml

@@ -1,22 +1,22 @@
-# We need a persistent volume for Vector to collec the logs, which will then be accessible 
+# We need a persistent volume for Vector to collect the logs, which will then be accessible
 # to the log parser.
 
 apiVersion: v1
 kind: PersistentVolume
 metadata:
-  name: codex-workflow-logs-pv
+  name: vector-logs-pv
 spec:
   capacity:
     storage: 10Gi
   accessModes:
     - ReadWriteMany
   hostPath:
-    path: /mnt/codex-workflow-logs
+    path: /mnt/vector-logs
 ---
 apiVersion: v1
 kind: PersistentVolumeClaim
 metadata:
-  name: codex-workflow-logs-pvc
+  name: vector-logs-pvc
 spec:
   accessModes:
     - ReadWriteMany