[Service/Component Name] Metrics¶

This document details the telemetry metrics exposed by the [Service or Component Name]. These metrics provide insights into the performance, reliability, and operational status of [briefly describe the component's purpose].

The primary meter for these metrics is [Your.Meter.Name].

Service Goal

[Provide a concise, one-sentence description of what this service is responsible for.]

Metrics¶

Use this table to list all metrics exposed by the component.

Metric Name	Type	Description
`metric_name_total`	Counter	`[Description of what the counter tracks, e.g., "Total number of operations attempted."]`
`metric_name_errors_total`	Counter	`[Description of what the error counter tracks.]`
`metric_name_duration_seconds`	Histogram	`[Description of the histogram, e.g., "The duration, in seconds, of each operation."]`
`metric_name_last_event_timestamp`	Observable Gauge	`[Description of the gauge, e.g., "The Unix epoch timestamp of the last successful event."]`
`metric_name_active_items`	Observable Gauge	`[Description of the gauge, e.g., "The number of items currently being processed."]`

Dimensions¶

Use this table to describe the dimensions (tags) associated with the metrics listed above.

Metric Name	Dimension Name	Possible Values	Description
`metric_name_total`	`operation`	`success`, `skipped`, `failure`	`[Description of the dimension.]`
	`reason`	`[value1]`, `[value2]`	`[Description of the dimension, e.g., "The reason an operation was skipped."]`
`metric_name_errors_total`	`error_type`	Exception Name	`[Description of the dimension.]`

Example KQL Queries¶

Provide useful KQL queries that users can adapt for their own monitoring and alerting needs in Azure Application Insights.

Generic Status BreakdownP95 DurationStale Event Alert

customMetrics
| where name == "metric_name_total"
| where timestamp > ago(1d)
| extend operation = tostring(customDimensions.operation)
| summarize Count = sum(value) by operation
| order by Count desc

customMetrics
| where name == "metric_name_duration_seconds"
| where timestamp > ago(7d)
| summarize percentiles(value, 95) by bin(timestamp, 1d)

customMetrics
| where name == "metric_name_last_event_timestamp"
| summarize LastEvent = max(value)
| extend AgeInSeconds = datetime_diff('second', now(), unixtime_seconds_todatetime(LastEvent))
| where AgeInSeconds > 3600 // 1 hour
| project readable_time_utc = unixtime_seconds_todatetime(LastEvent), AgeInSeconds

Metric Flow Diagram¶

Use Mermaid to create a simple diagram illustrating the process and showing where each metric is recorded.

graph TD
    A[Start] --> B{Decision Point};
    B -- Success --> C[Record: metric_name_total(success)];
    B -- Failure --> D[Record: metric_name_errors_total];