No details in Object Store Memory

adam · January 19, 2023, 2:18pm

How severe does this issue affect your experience of using Ray?

High: It blocks me to complete my task.

Explanation

The plot on the “Object Store Memory” in the dashboard (using Grafana and in Ray’s dashboard) doesn’t provide the separation for IN_MEMORY, UNSEALED, SPILLED, MAX. It has instead only VALUE and MAX. However I see the separation in e.g. Scheduler Task State (where I can see SUBMITED_TO_WORKER, RUNNING …)

ray 2.2.0
Python 3.8.16

Huaiwei_Sun · January 19, 2023, 7:46pm

I am able to see the location of the object store memory in ray dashboard (ray 2.2)…

Do you have a repro you could share that leads to the screenshot you have?
Could you also take a screenshot of this metric in ray dashboard UI? (I believe the one you showed here is from grafana)

adam · January 20, 2023, 2:37pm

Hi, thanks for the reply. I got the same image on the ray dashboard.
Let me give you some more details:

I run it on Mac (I checked on both Inter and M1 version, the problem exist on both)
I installed grafana and prometheus via brew and overwrote the configs using the files generated by ray in /tmp/ray/latest_session/ (I shortened the refresh rates for the example below, but that shouldn’t matter)
I have ray[default] installed.

Here is a short code that reproduces the issue:

import ray
import numpy as np
import time


@ray.remote
def f1(a):
    time.sleep(10)
    return a


if __name__ == "__main__":
    ray.init()
    sth_in_memory_id = ray.put(np.ones((1_000_000, 1)))
    f1_id = f1.remote(np.zeros((1_000_000, 1)))
    f1_res = ray.get(f1_id)
    input("Type enter when done")
    ray.shutdown()

Let me know if I can upload any logs that might help (and where to find them).

Huaiwei_Sun · January 20, 2023, 4:34pm

cc: @sangcho @rickyyx @aguo for ideas

rickyyx · January 20, 2023, 6:00pm

Hey @adam could you copy paste the grafana query of this chart? If you go to the grafana dashboard and click on editing the tab specifically?

adam · January 23, 2023, 3:25pm

Hi, I don’t think I can open the dashboard in edit mode. But I can download the dashboard.
I cannot enclose a json file. I’ll paste it in the next messages.

adam · January 23, 2023, 3:28pm

{
“__inputs”: [
{
“name”: “DS_PROMETHEUS”,
“label”: “Prometheus”,
“description”: “”,
“type”: “datasource”,
“pluginId”: “prometheus”,
“pluginName”: “Prometheus”
}
],
“__elements”: {},
“__requires”: [
{
“type”: “grafana”,
“id”: “grafana”,
“name”: “Grafana”,
“version”: “9.3.2”
},
{
“type”: “panel”,
“id”: “graph”,
“name”: “Graph (old)”,
“version”: “”
},
{
“type”: “datasource”,
“id”: “prometheus”,
“name”: “Prometheus”,
“version”: “1.0.0”
}
],
“annotations”: {
“list”: [
{
“builtIn”: 1,
“datasource”: {
“type”: “datasource”,
“uid”: “grafana”
},
“enable”: true,
“hide”: true,
“iconColor”: “rgba(0, 211, 255, 1)”,
“name”: “Annotations & Alerts”,
“target”: {
“limit”: 100,
“matchAny”: false,
“tags”: ,
“type”: “dashboard”
},
“type”: “dashboard”
}
]
},
“editable”: true,
“fiscalYearStartMonth”: 0,
“graphTooltip”: 0,
“id”: null,
“links”: ,
“liveNow”: false,
“panels”: [
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of tasks in a particular state.\n\nState: the task state, as described by rpc::TaskState proto in common.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 0
},
“hiddenSeries”: false,
“id”: 26,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(max_over_time(ray_tasks{State=~"FINISHED|FAILED",SessionName="$SessionName"}[14d])) by (State) or clamp_min(sum(ray_tasks{State!~"FINISHED|FAILED",SessionName="$SessionName"}) by (State), 0)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Task State”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “tasks”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of (live) tasks with a particular name.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 0
},
“hiddenSeries”: false,
“id”: 35,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_tasks{State!~"FINISHED|FAILED",SessionName="$SessionName"}) by (Name)”,
“interval”: “”,
“legendFormat”: “{{Name}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Active Tasks by Name”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “tasks”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of actors in a particular state.\n\nState: the actor state, as described by rpc::ActorTableData proto in gcs.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 8
},
“hiddenSeries”: false,
“id”: 33,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_actors{SessionName="$SessionName"}) by (State)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Actor State”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “actors”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of (live) actors with a particular name.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 8
},
“hiddenSeries”: false,
“id”: 36,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_actors{State!="DEAD",SessionName="$SessionName"}) by (Name)”,
“interval”: “”,
“legendFormat”: “{{Name}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Active Actors by Name”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “actors”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Logical CPU usage of Ray. The dotted line indicates the total number of CPUs. The logical CPU is allocated by num_cpus arguments from tasks and actors.\n\nNOTE: Ray’s logical CPU is different from physical CPU usage. Ray’s logical CPU is allocated by num_cpus arguments.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 16
},
“hiddenSeries”: false,
“id”: 27,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="CPU",State="USED",SessionName="$SessionName"}) by (instance)”,
“interval”: “”,
“legendFormat”: “CPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="CPU",SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler CPUs (logical slots)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “cores”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Object store memory usage by location. The dotted line indicates the object store memory capacity.\n\nLocation: where the memory was allocated, which is MMAP_SHM or MMAP_DISK to indicate memory-mapped page, SPILLED to indicate spillage to disk, and WORKER_HEAP for objects small enough to be inlined in worker memory. Refer to metric_defs.cc for more information.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 16
},
“hiddenSeries”: false,
“id”: 29,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_object_store_memory{SessionName="$SessionName"} / 1e9) by (Location)”,
“interval”: “”,
“legendFormat”: “{{Location}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="object_store_memory",SessionName="$SessionName"} / 1e9)”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Object Store Memory”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “gbytes”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: "Logical GPU usage of Ray. The dotted line indicates the total number of GPUs. The logical GPU is allocated by num_gpus arguments from tasks and actors. ",
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 24
},
“hiddenSeries”: false,
“id”: 28,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “ray_resources{Name="GPU",State="USED",SessionName="$SessionName"}”,
“interval”: “”,
“legendFormat”: “GPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="GPU",SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler GPUs (logical slots)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “GPUs”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of placement groups in a particular state.\n\nState: the placement group state, as described by the rpc::PlacementGroupTable proto in gcs.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 24
},
“hiddenSeries”: false,
“id”: 40,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_placement_groups{SessionName="$SessionName"}) by (State)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Placement Groups”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “placement groups”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 32
},
“hiddenSeries”: false,
“id”: 2,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “7.5.17”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “ray_node_cpu_utilization{instance=~"$Instance",SessionName="$SessionName"} * ray_node_cpu_count{instance=~"$Instance",SessionName="$SessionName"} / 100”,
“interval”: “”,
“legendFormat”: “CPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_node_cpu_count{SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Node CPU (hardware utilization)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “cores”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: "Node’s physical (hardware) GPU usage. The dotted line means the total number of hardware GPUs from the cluster. ",
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 32
},
“hiddenSeries”: false,
“id”: 8,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “7.5.17”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “ray_node_gpus_utilization{instance=~"$Instance",SessionName="$SessionName"} / 100”,
“interval”: “”,
“legendFormat”: “GPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_node_gpus_available{SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Node GPU (hardware utilization)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “GPUs”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: "Node’s physical (hardware) disk usage. The dotted line means the total amount of disk space from the cluster.\n\nNOTE: When Ray is deployed within a container, this shows the disk usage from the host machine. ",
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 40
},
“hiddenSeries”: false,
“id”: 6,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},

adam · January 23, 2023, 3:29pm

{
“__inputs”: [
{
“name”: “DS_PROMETHEUS”,
“label”: “Prometheus”,
“description”: “”,
“type”: “datasource”,
“pluginId”: “prometheus”,
“pluginName”: “Prometheus”
}
],
“__elements”: {},
“__requires”: [
{
“type”: “grafana”,
“id”: “grafana”,
“name”: “Grafana”,
“version”: “9.3.2”
},
{
“type”: “panel”,
“id”: “graph”,
“name”: “Graph (old)”,
“version”: “”
},
{
“type”: “datasource”,
“id”: “prometheus”,
“name”: “Prometheus”,
“version”: “1.0.0”
}
],
“annotations”: {
“list”: [
{
“builtIn”: 1,
“datasource”: {
“type”: “datasource”,
“uid”: “grafana”
},
“enable”: true,
“hide”: true,
“iconColor”: “rgba(0, 211, 255, 1)”,
“name”: “Annotations & Alerts”,
“target”: {
“limit”: 100,
“matchAny”: false,
“tags”: ,
“type”: “dashboard”
},
“type”: “dashboard”
}
]
},
“editable”: true,
“fiscalYearStartMonth”: 0,
“graphTooltip”: 0,
“id”: null,
“links”: ,
“liveNow”: false,
“panels”: [
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of tasks in a particular state.\n\nState: the task state, as described by rpc::TaskState proto in common.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 0
},
“hiddenSeries”: false,
“id”: 26,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(max_over_time(ray_tasks{State=~"FINISHED|FAILED",SessionName="$SessionName"}[14d])) by (State) or clamp_min(sum(ray_tasks{State!~"FINISHED|FAILED",SessionName="$SessionName"}) by (State), 0)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Task State”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “tasks”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of (live) tasks with a particular name.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 0
},
“hiddenSeries”: false,
“id”: 35,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_tasks{State!~"FINISHED|FAILED",SessionName="$SessionName"}) by (Name)”,
“interval”: “”,
“legendFormat”: “{{Name}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Active Tasks by Name”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “tasks”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of actors in a particular state.\n\nState: the actor state, as described by rpc::ActorTableData proto in gcs.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 8
},
“hiddenSeries”: false,
“id”: 33,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_actors{SessionName="$SessionName"}) by (State)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Actor State”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “actors”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of (live) actors with a particular name.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 8
},
“hiddenSeries”: false,
“id”: 36,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_actors{State!="DEAD",SessionName="$SessionName"}) by (Name)”,
“interval”: “”,
“legendFormat”: “{{Name}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Active Actors by Name”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “actors”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Logical CPU usage of Ray. The dotted line indicates the total number of CPUs. The logical CPU is allocated by num_cpus arguments from tasks and actors.\n\nNOTE: Ray’s logical CPU is different from physical CPU usage. Ray’s logical CPU is allocated by num_cpus arguments.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 16
},
“hiddenSeries”: false,
“id”: 27,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="CPU",State="USED",SessionName="$SessionName"}) by (instance)”,
“interval”: “”,
“legendFormat”: “CPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="CPU",SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler CPUs (logical slots)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “cores”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Object store memory usage by location. The dotted line indicates the object store memory capacity.\n\nLocation: where the memory was allocated, which is MMAP_SHM or MMAP_DISK to indicate memory-mapped page, SPILLED to indicate spillage to disk, and WORKER_HEAP for objects small enough to be inlined in worker memory. Refer to metric_defs.cc for more information.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 16
},
“hiddenSeries”: false,
“id”: 29,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_object_store_memory{SessionName="$SessionName"} / 1e9) by (Location)”,
“interval”: “”,
“legendFormat”: “{{Location}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="object_store_memory",SessionName="$SessionName"} / 1e9)”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Object Store Memory”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “gbytes”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: "Logical GPU usage of Ray. The dotted line indicates the total number of GPUs. The logical GPU is allocated by num_gpus arguments from tasks and actors. ",
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 24
},
“hiddenSeries”: false,
“id”: 28,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “ray_resources{Name="GPU",State="USED",SessionName="$SessionName"}”,
“interval”: “”,
“legendFormat”: “GPU Usage: {{instance}}”,
“queryType”: “randomWalk”,
“refId”: “A”
},
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_resources{Name="GPU",SessionName="$SessionName"})”,
“interval”: “”,
“legendFormat”: “MAX”,
“queryType”: “randomWalk”,
“refId”: “B”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler GPUs (logical slots)”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “GPUs”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “Current number of placement groups in a particular state.\n\nState: the placement group state, as described by the rpc::PlacementGroupTable proto in gcs.proto.”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 12,
“y”: 24
},
“hiddenSeries”: false,
“id”: 40,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “9.3.2”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],
“spaceLength”: 10,
“stack”: true,
“steppedLine”: false,
“targets”: [
{
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“exemplar”: true,
“expr”: “sum(ray_placement_groups{SessionName="$SessionName"}) by (State)”,
“interval”: “”,
“legendFormat”: “{{State}}”,
“queryType”: “randomWalk”,
“refId”: “A”
}
],
“thresholds”: ,
“timeRegions”: ,
“title”: “Scheduler Placement Groups”,
“tooltip”: {
“shared”: true,
“sort”: 0,
“value_type”: “individual”
},
“type”: “graph”,
“xaxis”: {
“mode”: “time”,
“show”: true,
“values”:
},
“yaxes”: [
{
“$$hashKey”: “object:628”,
“format”: “placement groups”,
“label”: “”,
“logBase”: 1,
“min”: “0”,
“show”: true
},
{
“$$hashKey”: “object:629”,
“format”: “short”,
“logBase”: 1,
“show”: true
}
],
“yaxis”: {
“align”: false
}
},
{
“aliasColors”: {},
“bars”: false,
“dashLength”: 10,
“dashes”: false,
“datasource”: {
“type”: “prometheus”,
“uid”: “${DS_PROMETHEUS}”
},
“description”: “”,
“fill”: 10,
“fillGradient”: 0,
“gridPos”: {
“h”: 8,
“w”: 12,
“x”: 0,
“y”: 32
},
“hiddenSeries”: false,
“id”: 2,
“legend”: {
“alignAsTable”: true,
“avg”: false,
“current”: true,
“hideEmpty”: false,
“hideZero”: true,
“max”: false,
“min”: false,
“rightSide”: false,
“show”: true,
“sort”: “current”,
“sortDesc”: true,
“total”: false,
“values”: true
},
“lines”: true,
“linewidth”: 1,
“nullPointMode”: “null”,
“options”: {
“alertThreshold”: true
},
“percentage”: false,
“pluginVersion”: “7.5.17”,
“pointradius”: 2,
“points”: false,
“renderer”: “flot”,
“seriesOverrides”: [
{
“$$hashKey”: “object:2987”,
“alias”: “MAX”,
“color”: “#1F60C4”,
“dashes”: true,
“fill”: 0,
“stack”: false
},
{
“$$hashKey”: “object:78”,
“alias”: “/FINISHED|FAILED|DEAD|REMOVED/”,
“hiddenSeries”: true
}
],

rickyyx · January 23, 2023, 6:44pm

Awesome, that will work.

Basically I just wanna make sure the query is correct.

Also, could you share your prometheus and grafana versions so we could repro as well? Thanks.

adam · January 24, 2023, 3:35pm

The messages with the dashboards were classified as spam. I’m sending a github’s gist of the Store Memory dashboard ray-dashboard-debug-Object Store Memory · GitHub
Versions:

grafana 9.3.2
prometheus 2.41.0

rickyyx · January 30, 2023, 6:04pm

Thanks @adam - will compare it with the config I have at my end ASAP - sorry for the delay.

rickyyx · February 15, 2023, 11:04pm

Sorry for the delay.

I believe the issue was this in the config you pasted:

Could you try editing the query explicitly like this :

Topic		Replies	Views
Does memory usage on ray dashboard include object store usage? Dashboard	3	694	January 31, 2023
`ray.cluster_resources()` output Ray Core	2	246	April 29, 2021
How to configure Memory monitor and Object Spilling? Ray Core	1	407	February 13, 2023
Memory leakage in object store memory Ray Core	3	99	February 29, 2024
Ray Spilled the object even though lots of /dev/shm is empty Ray Core	6	211	September 29, 2023

No details in Object Store Memory

Related Topics