Press CTRL+K to open search

AI metrics advisor

  • 4 minute(s) read
Prev Next

Overview

The AI metrics advisor is an AI Agent that provides best-practice metric monitoring recommendations for any resource within your Business Application. Instead of manually identifying which metrics to track, the agent analyzes the resource type and surfaces a prioritized list of metrics based on Azure SRE guidance, with recommended conditions, thresholds, and aggregation types ready to apply.

Business value

Setting up meaningful metric monitoring is time-consuming when done from scratch, and the wrong thresholds lead to either alert fatigue or missed incidents. The AI metrics advisor removes that guesswork by generating resource-type-specific recommendations pre-populated with Azure SRE best practices, allowing teams to apply a reliable monitoring baseline in a single step rather than configuring each metric manually.

How it works

  • When triggered, the agent evaluates the resource type and generates a ranked list of metric recommendations categorized as Critical, Important, or Optional.

  • Each recommendation includes the metric name, its purpose, suggested warning and error thresholds, condition, and aggregation type, giving full context for why each metric matters before applying it.

  • Recommendations can be selectively applied to the resource's monitoring configuration with a single click.

Recommendations

Each recommendation in the panel includes the following:

  • Metric: The name and unit of the metric
  • Recommendation: Priority level: Critical, Important, or Optional
  • Purpose: A plain-language explanation of what the metric tracks and why it matters
  • Condition: The comparison operator for the threshold rule
  • Warning / Error: Suggested threshold values for warning and error states. When a monitoring rule for the metric already exists with user-defined thresholds, the existing values are shown with a strikethrough alongside the AI-recommended values.
  • Aggregation: The aggregation type to use when evaluating the metric

Keep current values: For selected metrics that already have a monitoring rule with user-defined thresholds, a Keep current values option appears below the metric name. Clicking it retains the existing configured thresholds instead of applying the AI-recommended values.

Select the desired metrics and click Apply to add the selected recommendations to the resource's monitoring configuration.

Accessing AI metrics advisor

AI metrics advisor is accessible from three places:

Resource Monitoring tab: Open any resource within a Business Application, navigate to the Monitoring tab, and click AI metrics advisor in the Metrics section. The agent generates recommendations specific to that resource and allows thresholds to be applied directly to its monitoring configuration.

Monitoring profile (Create/Edit): While creating or editing a monitoring profile, select a resource type and click AI metrics advisor in the Metrics section. Recommendations are generated for the selected resource type and can be applied directly to the profile.

Create Business Application wizard: During Business Application creation, on the Monitoring rules step, select Configure monitoring rules, choose a resource type, and click AI metrics advisor in the Metrics section. Recommendations are generated for the selected resource type and can be applied directly before completing the wizard.

Troubleshooting

  1. AI metrics advisor button is not visible in the Metrics section
    Cause: The user has a Custom Role that does not have the AI Agents feature permission enabled.
    Fix: Go to User management > Roles, edit the relevant Custom Role, and set the AI Agents permission to Enabled under Specify feature permissions.

  2. AI metrics advisor generates no output on Private Hosting
    Cause: The Azure OpenAI connection has not been configured, or the configured endpoint or key is invalid.
    Fix: Go to Settings and verify your Azure OpenAI connection details. See OpenAI setup in Turbo360 for the full setup procedure.

  3. Recommendations appear but the Apply button does nothing
    Cause: No metrics are selected in the panel. The Apply button only activates when at least one metric is checked.
    Fix: Select one or more metrics from the recommendations list before clicking Apply.

  4. AI-recommended thresholds seem too low or too high for this resource
    Cause: Recommendations are generated based on Azure SRE best practices for the resource type, not the specific workload or traffic patterns of this particular resource.
    Fix: Review the suggested thresholds against your resource's actual usage before applying. Use the Keep current values option to retain existing thresholds for metrics where your configured values better reflect real workload behaviour.

  5. A metric I expected to see is not in the recommendations list
    Cause: The AI metrics advisor generates recommendations based on Azure SRE guidance for the selected resource type. Not all available metrics for a resource type may be included if they are not commonly recommended for monitoring.
    Fix: Apply the recommended metrics first, then configure any additional metrics manually from the Metrics section.

FAQs

  1. Does the AI metrics advisor work for all resource types?
    It works for all resource types supported in Business Applications. Recommendations are scoped to the specific resource type selected, so the metrics and thresholds generated are relevant to that type.

  2. Will applying recommendations overwrite my existing thresholds?
    Only if you choose to. For metrics that already have a monitoring rule with user-defined thresholds, the existing values are shown with a strikethrough alongside the AI-recommended values. Select Keep current values below the metric name to retain your configured thresholds and skip the AI-recommended ones.

  3. Are the recommended thresholds based on this specific resource's actual traffic?
    No. Recommendations are based on Azure SRE best practices for the resource type, not the individual resource's historical usage. Review them against your actual workload and adjust if needed after applying.

  4. Can I use AI metrics advisor when setting up monitoring for the first time?
    Yes. It is available during Business Application creation on the Monitoring rules step, so you can apply a recommended monitoring baseline before the Business Application is fully set up.

  5. Does using AI metrics advisor require an OpenAI connection on SaaS?
    No. SaaS deployments use Turbo360's in-house Azure OpenAI connection and require no additional setup. For Private Hosting deployments, you must configure your own Azure OpenAI resource under Settings before the agent can generate output.

Related articles