Extracting sum and count metrics from histogram

This documentation explains how to extract summary metrics (sum and count) from histogram-type metrics while excluding the original bucket data. This transformation is particularly useful for calculating averages.

Overview

The transformation configuration allows you to:

1. Extract count and sum metrics from histogram data
2. Filter out the original histogram buckets
3. Use the resulting metrics for average calculations

transform:
  error_mode: ignore
  metric_statements:
    - context:
        metric:
          conditions:
            - name == "myApp_latency_bucket"
          statements:
            - extract_count_metric(true)
            - extract_sum_metric(true)
  filter:
    metrics:
      exclude:
        match_type: strict
        metric_names:
          - myApp_latency_bucket

Generated Metrics

The transformation will create two new metrics:

1. myApp_latency_bucket_count: Total number of observations
2. myApp_latency_bucket_sum: Sum of all observations