You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: articles/advisor/azure-advisor-score.md
+35-30Lines changed: 35 additions & 30 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,14 +1,14 @@
1
1
---
2
2
title: Use Advisor score
3
-
description: Use Azure Advisor score to get the most out of Azure.
3
+
description: Use Azure Advisor score to measure optimization progress.
4
4
ms.topic: article
5
-
ms.date: 09/09/2020
5
+
ms.date: 07/12/2024
6
6
7
7
---
8
8
9
9
# Use Advisor score
10
10
11
-
## Introduction to Advisor score
11
+
## Introduction to score
12
12
13
13
Azure Advisor provides best practice recommendations for your workloads. These recommendations are personalized and actionable to help you:
14
14
@@ -18,7 +18,7 @@ Azure Advisor provides best practice recommendations for your workloads. These r
18
18
19
19
As a core feature of Advisor, Advisor score can help you achieve these goals effectively and efficiently.
20
20
21
-
To get the most out of Azure, it's crucial to understand where you are in your workload optimization journey. You need to know which services or resources are consumed well and which are not. Further, you'll want to know how to prioritize your actions, based on recommendations, to maximize the outcome.
21
+
To get the most out of Azure, it's crucial to understand where you are in your workload optimization journey. You need to know which services or resources are consumed well and which are not. Further, you want to know how to prioritize your actions, based on recommendations, to maximize the outcome.
22
22
23
23
It's also important to track and report the progress you're making in this optimization journey. With Advisor score, you can easily do all these things with the new gamification experience.
24
24
@@ -28,32 +28,40 @@ The Advisor score consists of an overall score, which can be further broken down
28
28
29
29
You can track the progress you make over time by viewing your overall score and category score with daily, weekly, and monthly trends. You can also set benchmarks to help you achieve your goals.
30
30
31
-
31
+
## Use Advisor score in the portal
32
+
33
+
1. Sign in to the [**Azure portal**](https://portal.azure.com).
34
+
35
+
1. Search for and select [**Advisor**](https://aka.ms/azureadvisordashboard) from any page.
36
+
37
+
1. Select **Advisor score** in the left menu pane to open score page.
38
+
39
+
:::image type="content" source="./media/advisor-score.png" alt-text="Screenshot that shows the Advisor Score entry point." lightbox="./media/advisor-score.png":::
32
40
33
41
## Interpret an Advisor score
34
42
35
43
Advisor displays your overall Advisor score and a breakdown for Advisor categories, in percentages. A score of 100% in any category means all your resources assessed by Advisor follow the best practices that Advisor recommends. On the other end of the spectrum, a score of 0% means that none of your resources assessed by Advisor follow Advisor's recommendations. Using these score grains, you can easily achieve the following flow:
36
44
37
45
***Advisor score** helps you baseline how your workload or subscriptions are doing based on an Advisor score. You can also see the historical trends to understand what your trend is.
38
-
***Score by category** for each recommendation tells you which outstanding recommendations will improve your score the most. These values reflect both the weight of the recommendation and the predicted ease of implementation. These factors help to make sure you can get the most value with your time. They also help you with prioritization.
46
+
***Score by category** for each recommendation tells you which outstanding recommendations improve your score the most. These values reflect both the weight of the recommendation and the predicted ease of implementation. These factors help to make sure you can get the most value with your time. They also help you with prioritization.
39
47
***Category score impact** for each recommendation helps you prioritize your remediation actions for each category.
40
48
41
49
The contribution of each recommendation to your category score is shown clearly on the **Advisor score** page in the Azure portal. You can increase each category score by the percentage point listed in the **Potential score increase** column. This value reflects both the weight of the recommendation within the category and the predicted ease of implementation to address the potentially easiest tasks. Focusing on the recommendations with the greatest score impact will help you make the most progress with time.
42
50
43
51
44
52
45
-
If any Advisor recommendations aren't relevant for an individual resource, you can postpone or dismiss those recommendations. They'll be excluded from the score calculation with the next refresh. Advisor will also use this input as additional feedback to improve the model.
53
+
If any Advisor recommendations aren't relevant for an individual resource, you can postpone or dismiss those recommendations. They'll be excluded from the score calculation with the next refresh. Advisor will also use this input as feedback to improve the model.
46
54
47
55
## How is an Advisor score calculated?
48
56
49
57
Advisor displays your category scores and your overall Advisor score as percentages. A score of 100% in any category means all your resources, *assessed by Advisor*, follow the best practices that Advisor recommends. On the other end of the spectrum, a score of 0% means that none of your resources, assessed by Advisor, follows Advisor recommendations.
50
58
51
-
**Each of the five categories has a highest potential score of 100.** Your overall Advisor score is calculated as a sum of each applicable category score, divided by the sum of the highest potential score from all applicable categories. For most subscriptions, that means Advisor adds up the score from each category and divides by 500. But *each category score is calculated only if you use resources that are assessed by Advisor*.
59
+
**Each of the five categories has a highest potential score of 100.** Your overall Advisor score is calculated as a sum of each applicable category score, divided by the sum of the highest potential score from all applicable categories. In most cases this means adding up five Advisor scores for each category and dividing by 500. But *each category score is calculated only if you use resources that are assessed by Advisor*.
52
60
53
61
### Advisor score calculation example
54
62
55
63
***Single subscription score:** This example is the simple mean of all Advisor category scores for your subscription. If the Advisor category scores are - **Cost** = 73, **Reliability** = 85, **Operational excellence** = 77, and **Performance** = 100, the Advisor score would be (73 + 85 + 77 + 100)/(4x100) = 0.84% or 84%.
56
-
***Multiple subscriptions score:** When multiple subscriptions are selected, the overall Advisor scores generated are weighted aggregate category scores. Here, each Advisor category score is aggregated based on resources consumed by subscriptions. After Advisor has the weighted aggregated category scores, Advisor does a simple mean calculation to give you an overall score for subscriptions.
64
+
***Multiple subscriptions score:** When multiple subscriptions are selected, the overall Advisor score is calculated as an average of aggregated category scores. Each category score is calculated using individual subscription score and subscription consumsumption based weight. Overall score is calculated as sum of aggregated category scores divided by the sum of the highest potential scores.
57
65
58
66
### Scoring methodology
59
67
@@ -62,39 +70,47 @@ The calculation of the Advisor score can be summarized in four steps:
62
70
1. Advisor calculates the *retail cost of impacted resources*. These resources are the ones in your subscriptions that have at least one recommendation in Advisor.
63
71
1. Advisor calculates the *retail cost of assessed resources*. These resources are the ones monitored by Advisor, whether they have any recommendations or not.
64
72
1. For each recommendation type, Advisor calculates the *healthy resource ratio*. This ratio is the retail cost of impacted resources divided by the retail cost of assessed resources.
65
-
1. Advisor applies three additional weights to the healthy resource ratio in each category:
73
+
1. Advisor applies three other weights to the healthy resource ratio in each category:
66
74
67
75
* Recommendations with greater impact are weighted heavier than recommendations with lower impact.
68
-
* Resources with long-standing recommendations will count more against your score.
76
+
* Resources with long-standing recommendations count more against your score.
69
77
* Resources that you postpone or dismiss in Advisor are removed from your score calculation entirely.
70
78
71
79
Advisor applies this model at an Advisor category level to give an Advisor score for each category. **Security** uses a [secure score](../defender-for-cloud/secure-score-security-controls.md) model. A simple average produces the final Advisor score.
72
80
73
-
## Advisor score FAQs
81
+
## Frequently Asked Questions (FAQs)
74
82
75
83
### How often is my score refreshed?
76
84
77
85
Your score is refreshed at least once per day.
78
86
87
+
### Why did my score change?
88
+
89
+
Your score can change if you remediate impacted resources by adopting the best practices that Advisor recommends. If you or anyone with permissions on your subscription has modified or created new resources, you might also see fluctuations in your score. Your score is based on a ratio of the cost-impacted resources relative to the total cost of all resources.
90
+
91
+
### I implemented a recommendation but my score did not change. Why the score did not increase?
92
+
93
+
The score does not reflect adopted recommendations right away. It takes at least 24 hours for the score to change after the recommendation is remediated.
94
+
79
95
### Why do some recommendations have the empty "-" value in the category score impact column?
80
96
81
97
Advisor doesn't immediately include new recommendations or recommendations with recent changes in the scoring model. After a short evaluation period, typically a few weeks, they're included in the score.
82
98
83
-
### Why is the Cost score impact greater for some recommendations even if they have lower potential savings?
99
+
### Why is the cost score impact greater for some recommendations even if they have lower potential savings?
84
100
85
-
Your **Cost** score reflects both your potential savings from underutilized resources and the predicted ease of implementing those recommendations. For example, extra weight is applied to impacted resources that have been idle for a longer time, even if the potential savings is lower.
101
+
Your **Cost** score reflects both your potential savings from underutilized resources and the predicted ease of implementing those recommendations. For example, extra weight is applied to impacted resources that have been idle for a long time, even if the potential savings are lower.
86
102
87
-
### Why don't I have a score for one or more categories or subscriptions?
103
+
### What does it mean when I see "Coming soon" in the score impact column?
88
104
89
-
Advisor generates a score only for the categories and subscriptions that have resources that are assessed by Advisor.
105
+
This message means that the recommendation is new, and we're working on bringing it to the Advisor score model. After this new recommendation is considered in a score calculation, you'll see the score impact value for your recommendation.
90
106
91
107
### What if a recommendation isn't relevant?
92
108
93
-
If you dismiss a recommendation from Advisor, it will be omitted from the calculation of your score. Dismissing recommendations also helps Advisor improve the quality of recommendations.
109
+
If you dismiss a recommendation from Advisor, it is excluded from the calculation of your score. Dismissing recommendations also helps Advisor improve the quality of recommendations.
94
110
95
-
### Why did my score change?
111
+
### Why don't I have a score for one or more categories or subscriptions?
96
112
97
-
Your score can change if you remediate impacted resources by adopting the best practices that Advisor recommends. If you or anyone with permissions on your subscription has modified or created new resources, you might also see fluctuations in your score. Your score is based on a ratio of the cost-impacted resources relative to the total cost of all resources.
113
+
Advisor generates a score only for the categories and subscriptions that have resources that are assessed by Advisor.
98
114
99
115
### How does Advisor calculate the retail cost of resources on a subscription?
100
116
@@ -112,21 +128,10 @@ No, not for now. But you can dismiss recommendations on individual resources if
112
128
113
129
The scoring methodology is designed to control for the number of resources on a subscription and service mix. Subscriptions with fewer resources can have higher or lower scores than subscriptions with more resources.
114
130
115
-
### What does it mean when I see "Coming soon" in the score impact column?
116
-
117
-
This message means that the recommendation is new, and we're working on bringing it to the Advisor score model. After this new recommendation is considered in a score calculation, you'll see the score impact value for your recommendation.
118
-
119
131
### Does my score depend on how much I spend on Azure?
120
132
121
133
No. Your score isn't necessarily a reflection of how much you spend. Unnecessary spending will result in a lower **Cost** score.
122
134
123
-
## Access Advisor Score
124
-
125
-
In the left pane, under the **Advisor** section, see **Advisor score**.
126
-
127
-
128
-
129
-
130
135
## Next steps
131
136
132
137
For more information about Advisor recommendations, see:
|`Azure OpenAI Requests`|HTTP|Count|Total number of calls made to the Azure OpenAI API over a period of time. Applies to PayGo, PTU, and PTU-managed SKUs.|`ApiName`, `ModelDeploymentName`,`ModelName`,`ModelVersion`, `OperationName`, `Region`, `StatusCode`, `StreamType`|
59
+
|`Active Tokens`| Usage | Total tokens minus cached tokens over a period of time. Applies to PTU and PTU-managed deployments. Use this metric to understand your TPS or TPM based utilization for PTUs and compare to your benchmarks for target TPS or TPM for your scenarios. |`ModelDeploymentName`,`ModelName`,`ModelVersion`|
59
60
|`Generated Completion Tokens`| Usage | Sum | Number of generated tokens (output) from an Azure OpenAI model. Applies to PayGo, PTU, and PTU-manged SKUs |`ApiName`, `ModelDeploymentName`,`ModelName`, `Region`|
60
61
|`Processed FineTuned Training Hours`| Usage |Sum| Number of training hours processed on an Azure OpenAI fine-tuned model. |`ApiName`, `ModelDeploymentName`,`ModelName`, `Region`|
61
62
|`Processed Inference Tokens`| Usage | Sum| Number of inference tokens processed by an Azure OpenAI model. Calculated as prompt tokens (input) + generated tokens. Applies to PayGo, PTU, and PTU-manged SKUs.|`ApiName`, `ModelDeploymentName`,`ModelName`, `Region`|
@@ -64,6 +65,7 @@ The following table summarizes the current subset of metrics available in Azure
64
65
|`Prompt Token Cache Match Rate`| HTTP | Average |**Provisioned-managed only**. The prompt token cache hit ration expressed as a percentage. |`ModelDeploymentName`, `ModelVersion`, `ModelName`, `Region`|
65
66
|`Time to Response`| HTTP | Average | Recommended latency (responsiveness) measure for streaming requests. **Applies to PTU, and PTU-managed deployments**. This metric does not apply to standard pay-go deployments. Calculated as time taken for the first response to appear after a user sends a prompt, as measured by the API gateway. This number increases as the prompt size increases and/or cache hit size reduces. Note: this metric is an approximation as measured latency is heavily dependent on multiple factors, including concurrent calls and overall workload pattern. In addition, it does not account for any client- side latency that may exist between your client and the API endpoint. Please refer to your own logging for optimal latency tracking.|`ModelDepIoymentName`, `ModelName`, and `ModelVersion`|
66
67
68
+
67
69
## Configure diagnostic settings
68
70
69
71
All of the metrics are exportable with [diagnostic settings in Azure Monitor](/azure/azure-monitor/essentials/diagnostic-settings). To analyze logs and metrics data with Azure Monitor Log Analytics queries, you need to configure diagnostic settings for your Azure OpenAI resource and your Log Analytics workspace.
Copy file name to clipboardExpand all lines: articles/ai-services/speech-service/speech-sdk.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -14,7 +14,7 @@ ms.author: eur
14
14
15
15
The Speech SDK (software development kit) exposes many of the [Speech service capabilities](overview.md), so you can develop speech-enabled applications. The Speech SDK is available [in many programming languages](quickstarts/setup-platform.md) and across platforms. The Speech SDK is ideal for both real-time and non-real-time scenarios, by using local devices, files, Azure Blob Storage, and input and output streams.
16
16
17
-
In some cases, you can't or shouldn't use the [Speech SDK](speech-sdk.md). In those cases, you can use REST APIs to access the Speech service. For example, use the [Speech to text REST API](rest-speech-to-text.md) for [batch transcription](batch-transcription.md) and [custom speech](custom-speech-overview.md).
17
+
In some cases, you can't or shouldn't use the [Speech SDK](speech-sdk.md). In those cases, you can use REST APIs to access the Speech service. For example, use the [Speech to text REST API](rest-speech-to-text.md) for [batch transcription](batch-transcription.md) and [custom speech](custom-speech-overview.md) model management.
0 commit comments