GetMetricDataV2 - "API contacts handled" metric, not giving the same results as in the Historical Metrics UI

What do you get if you change your filter setting from AGENT to the different CHANNEL that you have like CHAT (and others that you are trying to pull report for). Keep the groups value of AGENT and let us know.

Hello lautaroef,

I have re-replicated the GetMetricDataV2 - "API setup and initially encountered a scenario where the API returned no output, similar to your experience. However after further investigation , I managed to resolve the issue by making some additions to the code. This is the final edited code :

// Import required modules
import { ConnectClient, GetMetricDataV2Command } from "@aws-sdk/client-connect";

// Lambda handler function
export async function handler(event, context) {
    try {
        // Create AWS Connect client
        const connectClient = new ConnectClient({ region: "us-east-1" }); // Change the region as needed

        // Define input parameters for the GetMetricDataV2 command
        const metricsInput = {
            ResourceArn: "arn:aws:connect:us-east-1:<user_id>:instance/<same_instance_used_in_historical_metrics>",
            StartTime: new Date("2024-03-26T07:00:00.000Z"),
            EndTime: new Date("2024-03-28T06:59:59.999Z"),
            Interval: {
                TimeZone: 'America/Los_Angeles',
                IntervalPeriod: 'TOTAL',
            },
            Filters: [{ FilterKey: "AGENT", FilterValues: ['<AGENT ID>'] }], // agent ids
            Groupings: ["QUEUE"],
            Metrics: [
                // API Contacts Handled
                {
                    Name: 'CONTACTS_HANDLED',
                    MetricFilters: [
                        {
                            MetricFilterKey: 'INITIATION_METHOD',
                            MetricFilterValues: ['API'],
                        },
                    ],
                },
            ],
        };
       
        // Create GetMetricDataV2Command with the metrics input
        const command = new GetMetricDataV2Command(metricsInput);

        // Send the command to AWS Connect service and await response
        const response = await connectClient.send(command);

      // Log the response for debugging (optional)
        console.log(response);

        // Return response (or process it further)
        return response;

    } catch (error) {
        // Handle errors
        console.error("Error:", error);
        throw error; // Rethrow the error for Lambda to handle
    }
};

When I was running your original code, I noticed that your code had the line Name: 'CONTACTS_HANDLED' placed inside the Metrics array, where metric filters are typically specified. However, this caused confusion as it implied that 'CONTACTS_HANDLED' was a filter, when its actual purpose was to represent the metric we wanted to retrieve. Recognizing this, I moved Name: 'CONTACTS_HANDLED' outside of the MetricFilters array in my modified version of the code, and this is the output I got:

Response { "$metadata": { "httpStatusCode": 200, "requestId": "<REQUEST ID>", "attempts": 1, "totalRetryDelay": 0 }, "MetricResults": [ { "Collections": [ { "Metric": { "MetricFilters": [ { "MetricFilterKey": "INITIATION_METHOD", "MetricFilterValues": [ "API" ], "Negate": false } ], "Name": "CONTACTS_HANDLED" }, "Value": 4 } ], "Dimensions": { "QUEUE": "<QUEUE ID>", "QUEUE_ARN": "QUEUE ARN" }, "MetricInterval": { "EndTime": "2024-03-28T07:00:00.000Z", "Interval": "TOTAL", "StartTime": "2024-03-26T07:00:00.000Z" } } ] }

I observed that the data retrieved correlated with the test calls I had previously placed via the API. This suggests that the code is now accurately fetching the desired metric data.