Skip to content

Same agent Smaller bill. The tools and infrastructure to get more work from your AI. Fewer tokens. Less compute. Lower cost.

[Fix]: gpu utilization keys updated (#453)

Aman Agarwal7a08f0b5f36b
diff --git a/docs/latest/features/metrics.mdx b/docs/latest/features/metrics.mdx
index 34760022..e1ba1805 100644
--- a/docs/latest/features/metrics.mdx
+++ b/docs/latest/features/metrics.mdx
@@ -82,10 +82,10 @@ This section outlines the OpenTelemetry metrics collected by **OpenLIT** from ap
| Metric Name | Description | Unit | Type | Attributes |
|-------------------------------------|------------------------------------------|------------|-------|------------------------------------------------------------------------------------------------------------|
-| `gpu.utilization_percentage` | GPU Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.enc.utilization_percentage` | GPU encoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.dec.utilization_percentage` | GPU decoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.temperature` | GPU Temperature in Celsius | `Celcius` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.utilization` | GPU Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.enc.utilization` | GPU encoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.dec.utilization` | GPU decoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.temperature` | GPU Temperature in Celsius | `Celsius` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.fan_speed` | GPU Fan Speed (0-100) as an integer | `Integer` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.memory.available` | Available GPU Memory in MB | `MB` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.memory.total` | Total GPU Memory in MB | `MB` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
diff --git a/otel-gpu-collector/README.md b/otel-gpu-collector/README.md
index 1b4e43e2..0af84899 100644
--- a/otel-gpu-collector/README.md
+++ b/otel-gpu-collector/README.md
@@ -78,10 +78,10 @@ For more details, check out the [OpenLIT documentation](https://docs.openlit.io/
| Metric Name | Description | Unit | Type | Attributes |
|-------------------------------------|------------------------------------------|------------|-------|------------------------------------------------------------------------------------------------------------|
-| `gpu.utilization_percentage` | GPU Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.enc.utilization_percentage` | GPU encoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.dec.utilization_percentage` | GPU decoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
-| `gpu.temperature` | GPU Temperature in Celsius | `Celcius` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.utilization` | GPU Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.enc.utilization` | GPU encoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.dec.utilization` | GPU decoder Utilization in percentage | `percent` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
+| `gpu.temperature` | GPU Temperature in Celsius | `Celsius` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.fan_speed` | GPU Fan Speed (0-100) as an integer | `Integer` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.memory.available` | Available GPU Memory in MB | `MB` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
| `gpu.memory.total` | Total GPU Memory in MB | `MB` | Gauge | `telemetry.sdk.name`, `gen_ai.application_name`, `gen_ai.environment`, `gpu_index`, `gpu_name`, `gpu_uuid` |
diff --git a/src/client/package.json b/src/client/package.json
index deae7030..222f49f9 100644
--- a/src/client/package.json
+++ b/src/client/package.json
@@ -1,6 +1,6 @@
{
"name": "openlit",
- "version": "1.9.1",
+ "version": "1.9.2",
"private": true,
"scripts": {
"dev": "next dev",
diff --git a/src/client/src/app/(playground)/dashboard/gpu/index.tsx b/src/client/src/app/(playground)/dashboard/gpu/index.tsx
index 3310ee0f..860453ce 100644
--- a/src/client/src/app/(playground)/dashboard/gpu/index.tsx
+++ b/src/client/src/app/(playground)/dashboard/gpu/index.tsx
@@ -7,11 +7,7 @@ export default function GPUDashboard() {
<NumberStats />
<div className="grid-cols-2 grid gap-4">
<GPUMetric
- chartKeys={[
- "utilization_percentage",
- "enc_utilization_percentage",
- "dec_utilization_percentage",
- ]}
+ chartKeys={["utilization", "enc_utilization", "dec_utilization"]}
url="/api/metrics/gpu/utilization/time"
title="Avg Utilization Percentage (%)"
/>
@@ -26,7 +22,7 @@ export default function GPUDashboard() {
url="/api/metrics/gpu/memory/time"
title="Memory (MB)"
/>
-
+
<GPUMetric
chartKeys={["power.limit", "power.draw"]}
url="/api/metrics/gpu/power/time"
diff --git a/src/client/src/app/(playground)/dashboard/gpu/number-stats.tsx b/src/client/src/app/(playground)/dashboard/gpu/number-stats.tsx
index 9cb7d065..fd752e63 100644
--- a/src/client/src/app/(playground)/dashboard/gpu/number-stats.tsx
+++ b/src/client/src/app/(playground)/dashboard/gpu/number-stats.tsx
@@ -5,7 +5,7 @@ function NumberStats() {
return (
<div className="grid gap-4 md:grid-cols-2 lg:grid-cols-4">
<StatCard
- dataKey="utilization_percentage"
+ dataKey="utilization"
heading="Avg utilization percentage"
textSuffix="%"
icon={UtilityPole}
diff --git a/src/client/src/app/api/metrics/gpu/utilization/average/route.ts b/src/client/src/app/api/metrics/gpu/utilization/average/route.ts
index 7785a1dd..cdf28ca0 100644
--- a/src/client/src/app/api/metrics/gpu/utilization/average/route.ts
+++ b/src/client/src/app/api/metrics/gpu/utilization/average/route.ts
@@ -15,7 +15,7 @@ export async function POST(request: Request) {
const validationParam = validateMetricsRequest(
params,
- validateMetricsRequestType.AVERAGE_UTILIZATION_PERCENTAGE
+ validateMetricsRequestType.AVERAGE_UTILIZATION
);
if (!validationParam.success)
diff --git a/src/client/src/app/api/metrics/gpu/utilization/time/route.ts b/src/client/src/app/api/metrics/gpu/utilization/time/route.ts
index 87d7d161..380aa2b9 100644
--- a/src/client/src/app/api/metrics/gpu/utilization/time/route.ts
+++ b/src/client/src/app/api/metrics/gpu/utilization/time/route.ts
@@ -15,7 +15,7 @@ export async function POST(request: Request) {
const validationParam = validateMetricsRequest(
params,
- validateMetricsRequestType.UTILIZATION_PERCENTAGE_PER_TIME
+ validateMetricsRequestType.UTILIZATION_PER_TIME
);
if (!validationParam.success)
diff --git a/src/client/src/helpers/platform.ts b/src/client/src/helpers/platform.ts
index c5c57ee2..4b75723f 100644
--- a/src/client/src/helpers/platform.ts
+++ b/src/client/src/helpers/platform.ts
@@ -45,8 +45,8 @@ export const validateMetricsRequestType = {
// fanspeed
FANSPEED_PER_TIME: "FANSPEED_PER_TIME",
// utilization
- AVERAGE_UTILIZATION_PERCENTAGE: "AVERAGE_UTILIZATION_PERCENTAGE",
- UTILIZATION_PERCENTAGE_PER_TIME: "UTILIZATION_PERCENTAGE_PER_TIME",
+ AVERAGE_UTILIZATION: "AVERAGE_UTILIZATION",
+ UTILIZATION_PER_TIME: "UTILIZATION_PER_TIME",
// Vectordb
GENERATION_BY_OPERATION: "GENERATION_BY_OPERATION",
diff --git a/src/client/src/lib/platform/common.ts b/src/client/src/lib/platform/common.ts
index 1ba33e8b..a239e068 100644
--- a/src/client/src/lib/platform/common.ts
+++ b/src/client/src/lib/platform/common.ts
@@ -31,9 +31,9 @@ export interface MetricParams {
}
export type GPU_TYPE_KEY =
- | "utilization_percentage"
- | "enc.utilization_percentage"
- | "dec.utilization_percentage"
+ | "utilization"
+ | "enc.utilization"
+ | "dec.utilization"
| "temperature"
| "fan_speed"
| "memory.available"
diff --git a/src/client/src/lib/platform/gpu/utilization.ts b/src/client/src/lib/platform/gpu/utilization.ts
index 6dd8df3c..12a68b97 100644
--- a/src/client/src/lib/platform/gpu/utilization.ts
+++ b/src/client/src/lib/platform/gpu/utilization.ts
@@ -11,10 +11,12 @@ import {
export async function getAverageUtilization(params: GPUMetricParams) {
const query = `
SELECT
- ROUND(AVG(Value), 2) as utilization_percentage
+ ROUND(AVG(Value), 2) as utilization
FROM
${OTEL_GPUS_TABLE_NAME}
- WHERE ${getFilterWhereConditionForGPU(params)} AND MetricName = 'gpu.utilization_percentage'
+ WHERE ${getFilterWhereConditionForGPU(
+ params
+ )} AND MetricName = 'gpu.utilization'
GROUP BY
MetricName
`;
@@ -23,11 +25,7 @@ export async function getAverageUtilization(params: GPUMetricParams) {
}
export async function getUtilizationParamsPerTime(params: GPUMetricParams) {
- const keys = [
- "utilization_percentage",
- "enc.utilization_percentage",
- "dec.utilization_percentage",
- ];
+ const keys = ["utilization", "enc.utilization", "dec.utilization"];
const { start, end } = params.timeLimit;
const dateTrunc = dateTruncGroupingLogic(end as Date, start as Date);