HTTP Requests by Colo Groups to HTTP Requests by Adaptive Groups

This guide shares considerations when migrating from the deprecated httpRequests1mByColoGroups and httpRequests1dByColoGroups GraphQL API nodes to the httpRequestsAdaptiveGroups GraphQL API node.

For example, if you wanted to see which five data centers had the most number of requests, the total number of those requests, and the total amount of data transfer, in the past you used the httpRequests1mByColoGroups GraphQL API node as in the following example:

{
	viewer {
		zones(filter: { zoneTag: $zoneTag }) {
			series: httpRequests1mByColoGroups(
				limit: 5
				orderBy: [sum_requests_DESC]
				filter: { datetime_geq: $start, datetime_lt: $end }
			) {
				sum {
					requests
					bytes
				}
				dimensions {
					coloCode
				}
			}
		}
	}
}

Response

{
	"data": {
		"viewer": {
			"zones": [
				{
					"series": [
						{
							"dimensions": {
								"coloCode": "LHR"
							},
							"sum": {
								"bytes": 18260055,
								"requests": 4404
							}
						},
						{
							"dimensions": {
								"coloCode": "AMS"
							},
							"sum": {
								"bytes": 17563009,
								"requests": 4302
							}
						},
						{
							"dimensions": {
								"coloCode": "CDG"
							},
							"sum": {
								"bytes": 17200434,
								"requests": 4032
							}
						},
						{
							"dimensions": {
								"coloCode": "PTY"
							},
							"sum": {
								"bytes": 10400209,
								"requests": 2707
							}
						},
						{
							"dimensions": {
								"coloCode": "JIB"
							},
							"sum": {
								"bytes": 9040105,
								"requests": 2601
							}
						}
					]
				}
			]
		}
	},
	"errors": null
}

With the deprecation of the httpRequests1mByColoGroups and httpRequests1dByColoGroups GraphQL API nodes, use the httpRequestsAdaptiveGroups GraphQL API node to access the same data (count, sum(edgeResponseBytes), and visits).

Request

query MigrationSample($zoneTag: string, $start: Time, $end: Time) {
	viewer {
		zones(filter: { zoneTag: $zoneTag }) {
			series: httpRequestsAdaptiveGroups(
				limit: 5
				orderBy: [count_DESC]
				filter: {
					datetime_geq: $start
					datetime_lt: $end
					requestSource: "eyeball"
				}
			) {
				count
				avg {
					sampleInterval
				}
				sum {
					visits
					edgeResponseBytes
				}
				dimensions {
					coloCode
				}
			}
		}
	}
}

Response

{
	"data": {
		"viewer": {
			"zones": [
				{
					"series": [
						{
							"avg": {
								"sampleInterval": 10
							},
							"count": 4350,
							"dimensions": {
								"coloCode": "LHR"
							},
							"sum": {
								"edgeResponseBytes": 17860000,
								"visits": 4120
							}
						},
						{
							"avg": {
								"sampleInterval": 10
							},
							"count": 4210,
							"dimensions": {
								"coloCode": "AMS"
							},
							"sum": {
								"edgeResponseBytes": 17110000,
								"visits": 3910
							}
						},
						{
							"avg": {
								"sampleInterval": 10
							},
							"count": 3890,
							"dimensions": {
								"coloCode": "CDG"
							},
							"sum": {
								"edgeResponseBytes": 17050000,
								"visits": 3700
							}
						},
						{
							"avg": {
								"sampleInterval": 10
							},
							"count": 2550,
							"dimensions": {
								"coloCode": "PTY"
							},
							"sum": {
								"edgeResponseBytes": 10286000,
								"visits": 2130
							}
						},
						{
							"avg": {
								"sampleInterval": 10
							},
							"count": 2410,
							"dimensions": {
								"coloCode": "JIB"
							},
							"sum": {
								"edgeResponseBytes": 9029000,
								"visits": 2080
							}
						}
					]
				}
			]
		}
	},
	"errors": null
}

This query says:

• Given the indicated zones, limit, and time range.
• Fetch the total number of requests (as count), the total amount of data transfer (as edgeResponseBytes of sum object), and the total number of visits per data center.

A few points to note:

• Adding the requestSource filter for eyeball returns request, data transfer, and visit data about only the end users of your website.
• Instead of requests, the httpRequestsAdaptiveGroups node reports count, which indicates the number of requests per data center.
• To measure data transfer, use sum(edgeResponseBytes). Note that in the old API this was called bandwidth even though it actually measured data transfer.
• unique visitors per colocation is not supported in httpRequestsAdaptiveGroups, but the httpRequestsAdaptiveGroups API does support visits. A visit is defined as a page view that originated from a different website or direct link. Cloudflare checks where the HTTP referer does not match the hostname. One visit can consist of multiple page views.