​
Overview

Whenever changes occur in your ngrok account or when traffic transits through your endpoints, a log is generated. You may export these logs and in some cases, filter them to those relevant to you and publish them to any number of destinations. ngrok’s logging system was designed for three primary use cases:
  • Sending logs of your ngrok traffic to external services such as Datadog and Amazon CloudWatch Logs
  • Sending audit logs of ngrok configuration changes to your SIEM
  • Enabling you to programmatically respond to events on your ngrok account
There are over 40 log sources published via ngrok at the time of writing. They break down into two important classes of logs:
  • Traffic Logs: When traffic transits through your endpoints like processing an HTTP request or TCP connection
  • Audit Logs: Changes to your account like create/update/delete of Domains, API Keys, IP Policies, etc.

​
Concepts

ngrok’s logging system is composed of three simple primitives. You create a Log Export to subscribe to a set of Log Sources and publish those logs to one or more Log Destinations.
  • Log Exporting: Define which Log Sources to capture and which Log Destinations to publish to.
  • Log Sources: The ngrok log types to capture and optionally a set of fields to record and a boolean filter expression.
  • Log Destinations: Where logs are sent, e.g. Datadog, CloudWatch, Kinesis.

​
Quickstart

You can create your first Log Export on the Log Exporting page of your ngrok Dashboard to begin capturing and publishing logs. We also publish guides to get started with each of ngrok’s Log Destinations:

​
Log Exporting

A Log Export defines which Log Sources to capture and which destinations to publish to. If you’re familiar with other logging systems, they may call this a listener, a hook, a probe or a tap. When you create a Log Export, you select:
  • One or more Log Sources to subscribe to
  • One or more Log Destinations to publish the subscribed logs to
Log Exports can be created on the ngrok Dashboard or via the Log Exporting API Resource.

​
Log Sources

Log Sources choose which logs a Log Export captures. Other logging systems may call this a selector. The ngrok Log Sources reference documents the full list of Log Sources and their fields.

​
Versioning and Payload Stability

Log Sources include a version number in their name. Each Log Source ends with a version string after a dot, e.g. api_key_created.v0. New fields may be added to the payloads of a Log Source without a version change. Any other change to the log representation will lead to the introduction of a new version of the Log Source.

​
Audit Logs

ngrok emits Audit Logs when changes are made to your account like create/update/delete of Domains, API Keys, IP Policies, etc.

​
Traffic Logs

When traffic transits through your endpoints, ngrok emits Traffic Logs like processing an HTTP request or TCP connection. Traffic Logs support the ability to select exactly which fields you’d like to include in the captured log. This provides you with control over how much data is sent to the destinations. Traffic Logs also support the ability to be filtered. As part of the Log Source definition you can specify a CEL expression which is evaluated against each log to determine if it should be captured. See the Filters section for further detail.

​
Filters

You may specify a filter on Traffic Logs since the velocity of these logs can be quite high. Filters are a boolean expression defined in Google’s Common Expression Language (CEL). Filters are evaluated on each log as it is published to determine whether it should be sent to the Log Export’s destinations. Traffic Logs are often high cardinality. We support filters because you may only interested in logging a subset of traffic (e.g. specific client IPs). Filters also enable you to control spend on downstream destinations where you incur costs based on the quantity of data processed and stored. The log object is made available to the CEL filter expression as the ev variable. ev corresponds to the object field of the captured log. Use standard JSONPath syntax to access the fields of the log you want to use in your filters. Filter expressions have the following limitations:
  • Only some fields of an log are available in the ev object. The Log Source’s reference documentation will specify whether a field may be evaluated in a filter expression.
  • Log metadata fields like principal or event_timestamp are not available

​
Filter Examples

Filter for connections from a particular hostname
ev.conn.server_name.matches("ngrok-docs-examples\\.ngrok\\.dev")
Filter for connections to an endpoint that didn’t use HTTPS
ev.conn.server_name.matches(".*-your-org\\.ngrok\\.dev") &&
ev.conn.server_port == 80
Filter for connections to a hostname that exclude traffic from a client IP
ev.conn.client_ip != "2601:0:8200:0:4cd7:fd52:0:7823" &&
ev.conn.server_name == "ngrok-docs-examples.ngrok.dev"

​
Log Destinations

An Log Destination encapsulates the configuration to publish logs to other systems. Other logging systems may call this a sink or output. Log Destinations are typically third-party logging aggregators. The following destinations are currently supported: Each destination requires provider-specific configuration. If you create a destination in the ngrok dashboard, you’ll be prompted to send a test log to verify the integration.
When configuring AWS destinations you’ll be prompted to optionally download a small helper script which will automatically configure the appropriate IAM objects necessary for integration. You may also set these values up via the AWS Console or tools like Terraform or Pulumi.
Amazon S3 is not a directly supported destination. Instead, configure Amazon Firehose to deliver logs into an S3 bucket.

​
Log Payloads

Logs are serialized as JSON when they are published to a destination. All logs include the following fields:
NameDescriptionExample
account_idunique identifier for the account, always prefixed with ac_ac_2OtNvAlhso10Gx6s7eupzX3F98q
event_idunique identifier for this log, always prefixed with ev_ev_1vPlyBW3OR44bpPphS4HIZyajDD
event_typeidentifies the object, action, and version of the logip_policy_created.v0
event_timestamptimestamp of when the log fired in RFC 3339 format2021-07-16T21:44:37Z
objectthe log objectSee examples below
principalan object of the principal who emitted this log, null for traffic logsSee example below
The object property of the log is distinct for each Log Source and contain a JSON object with additional information about the log. For Audit Logs, the object representation is identical to its API resource at the time of capture. Traffic Logs define their own object representation because they have no corresponding API resource definition.

​
Principal Object

The principal object in every log describe the user or bot user responsible for initiating the log. Principal is defined for all Audit Logs and it is null for Traffic Logs (since they do not include a credential).
NameDescriptionExample
idunique identifier for the principal, either a user or botusr_2OtNv9qH5Nk4NuNeszZ39gBxZ4H
subjecthuman readable unique identifier for the principal, either a user email or a bot user namefoo@example.com
sourcewhere the principal initiated the log, either ‘Dashboard’ or ‘API’API
credentialid and uri resource for the credential used for authentication, null if source is ‘Dashboard’{"id": "ak_2Oxt94wYsBTLwFUoMZcJRvJTaub","uri": "https://api.ngrok.com/api_keys/ak_2Oxt94wYsBTLwFUoMZcJRvJTaub"}

​
Example Payloads

​
ip_policy_created.v0 log

{
	"event_id": "ev_25X2AsJ5xpvuOParTYUQWe12XKo",
	"event_type": "ip_policy_created.v0",
	"event_timestamp": "2022-02-23T23:29:29Z",
	"account_id": "ac_2OtNvAlhso10Gx6s7eupzX3F98q",
	"principal": {
		"id": "usr_2OtNv9qH5Nk4NuNeszZ39gBxZ4H",
		"subject": "foo@example.com",
		"source": "API",
		"credential": {
			"id": "ak_2Oxt94wYsBTLwFUoMZcJRvJTaub",
			"uri": "https://api.ngrok.com/api_keys/ak_2Oxt94wYsBTLwFUoMZcJRvJTaub"
		}
	},
	"object": {
		"id": "ipp_25X2Ao39z73FlVQKZ1iReMPe6Qv",
		"uri": "https://api.ngrok.com/ip_policies/ipp_25X2Ao39z73FlVQKZ1iReMPe6Qv",
		"created_at": "2022-02-23T23:29:29Z",
		"description": "Home network IP",
		"metadata": "",
		"action": "allow"
	}
}

​
http_request_complete.v0 log

{
	"event_id": "ev_25X3yFS6TDkig1KDJWIc4nnJO0c",
	"event_type": "http_request_complete.v0",
	"event_timestamp": "2022-02-23T23:44:16Z",
	"account_id": "ac_2OtNvAlhso10Gx6s7eupzX3F98q",
	"object": {
		"conn": {
			"client_ip": "2601:0:8200:9e:4cd7:0:c97f:7823",
			"server_name": "ngrok-docs-example.ngrok.app",
			"server_port": ""
		},
		"http": {
			"request": {
				"first_byte_ts": null,
				"last_byte_ts": null,
				"method": "GET",
				"url": {
					"path": "/docs/obs"
				},
				"version": "HTTP/2.0"
			},
			"response": {
				"body_length": 13079,
				"first_byte_ts": "2022-02-23T23:44:16.732791273Z",
				"last_byte_ts": "2022-02-23T23:44:16.737257209Z",
				"status_code": 200
			}
		}
	}
}

​
tcp_connection_closed.v0 log

{
	"event_id": "ev_25X4osod1q306srserDeFyghTC4",
	"event_type": "tcp_connection_closed.v0",
	"event_timestamp": "2022-02-23T23:51:14Z",
	"account_id": "ac_2OtNvAlhso10Gx6s7eupzX3F98q",
	"object": {
		"conn": {
			"bytes_in": 3437,
			"bytes_out": 90256,
			"client_ip": "2601:0:8200:9e:4cd7:0:c97f:7823",
			"end_ts": "2022-02-23T23:51:14.005372199Z",
			"server_name": "ngrok-docs-example.ngrok.app",
			"server_port": "",
			"start_ts": "2022-02-23T23:44:16.528374173Z"
		}
	}
}

​
Pricing

Log Exports are available to free and Pay-as-you-go plans. For more details, see our ngrok pricing page. The Traffic Inspector is included with all ngrok accounts with a retention period starting at 1 day and going up to 3 days. An additional 90 days can be purchased as an add-on to the Pay-as-you-go plan. For more details, see our ngrok pricing page.