This document defines a mechanism that enables developers to declare a network error reporting policy for a web application. A user agent can use this policy to report encountered network errors that prevented it from successfully fetching requested resources.

Introduction

Accurately measuring performance characteristics of web applications is an important aspect in helping site developers understand how to improve their web applications. The worst case scenario is the failure to load the application, or a particular resource, due to a network error, and to address such failures the developer requires assistance from the user agent to identify when, where, and why such failures are occurring.

Today, application developers do not have real-time web application availability data from their end users. For example, if the user fails to load the page due to a network error, such as a failed DNS lookup, a connection timeout, a reset connection, or other reasons, the site developer is unable to detect and address this issue. Note that these kinds of network errors cannot be detected purely server-side, since by definition the client might not have been able to successfully establish a connection with the server.

Existing methods (such as synthetic monitoring) provide a partial solution by placing monitoring nodes in predetermined geographic locations, but require additional infrastructure investments, and cannot provide truly global and near real-time availability data for real end users.

Network Error Logging (NEL) addresses this need by defining a mechanism enabling web applications to declare a reporting policy that can be used by the user agent to report network errors for a given origin. A web application opts into using NEL by supplying a NEL HTTP response header field that describes the desired NEL policy. This policy instructs the user agent to log information about requests to that origin, and to attempt to deliver that information to a group of endpoints previously configured using the Reporting API [[!REPORTING]]. As the name implies, NEL reports are primarily used to describe errors. However, in order to determine rates of errors across different client populations, we must also know how many successful requests are occurring; these successful requests can also be reported via the NEL mechanism.

For example, if the user agent fails to fetch a resource from https://www.example.com due to an aborted TCP connection, the user agent would queue the following report via the Reporting API:

type
"network-error"
endpoint group
the endpoint group configured by the report_to field
settings
TODO
data
{
  "referrer": "https://referrer.com/",
  "sampling_fraction": 1.0,
  "server_ip": "123.122.121.120",
  "protocol": "http/1.1",
  "elapsed_time": 321,
  "phase": "connection",
  "type": "tcp.aborted"
}
      

See for an explanation of the communicated fields and format of the report, and for more hands-on examples of NEL registration and reporting process.

Conformance requirements

All diagrams, examples, and notes in this specification are non-normative, as are all sections explicitly marked non-normative. Everything else in this specification is normative.

The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in the normative parts of this document are to be interpreted as described in [[!RFC2119]].

Requirements phrased in the imperative as part of algorithms (such as "strip any leading space characters" or "return false and abort these steps") are to be interpreted with the meaning of the key word ("must", "should", "may", etc) used in introducing the algorithm.

Some conformance requirements are phrased as requirements on attributes, methods or objects. Such requirements are to be interpreted as requirements on the user agent.

Conformance requirements phrased as algorithms or specific steps may be implemented in any manner, so long as the end result is equivalent. (In particular, the algorithms defined in this specification are intended to be easy to follow, and not intended to be performant.)

Dependencies

DNS

The following terms are defined in the DNS specification: [[!RFC1034]]

  • domain name
  • domain namespace tree
  • resolver
Fetch

The following terms are defined in the Fetch specification: [[!FETCH]]

  • client
  • CORS-preflight request
  • extract header list values
  • header list contains
  • header name
  • header value
  • HTTP-network fetch
  • HTTP-network-or-cache fetch
  • redirect status
  • request header list
  • response
  • response header list
  • response trailer
HSTS

The following terms are defined in the HSTS specification: [[!RFC6797]]

  • superdomain match
HTML

The following terms are defined in the HTML specification: [[!HTML]]

  • navigate
  • navigator.onLine
  • resource origin
HTTP

The following terms are defined in the HTTP specification: [[!RFC7230]], [[!RFC7231]], [[RFC7232]], [[!RFC7234]]

  • 200 status code
  • 4xx status code
  • 5xx status code
  • ETag
  • If-None-Match
  • local cache
  • persistent connections
  • request
  • request method
  • resource representation
  • response headers
  • server
  • status code
HTTP JSON field values

The following terms are defined in the HTTP-JFV specification: [[!HTTP-JFV]]

  • json-field-value
JSON

The following terms are defined in the JSON specification: [[!RFC7159]]

  • JSON object
Referrer Policy

The following terms are defined in the Referrer Policy specification: [[!REFERRER-POLICY]]

  • referrer policy
Reporting API

The following terms are defined in the Reporting API specification: [[!REPORTING]]

  • endpoint group
  • report
  • report body
  • report type
  • visible to ReportingObservers
Resource Timing

The following terms are defined in the Resource Timing specification: [[!RESOURCE-TIMING-2]]

  • network protocol
Secure Contexts

The following terms are defined in the Secure Contexts specification: [[!SECURE-CONTEXTS]]

  • the "Is origin potentially trustworthy?" algorithm
  • potentially trustworthy origin
URL

The following terms are defined in the URL specification: [[!URL]]

  • fragment
  • host
  • URL

Concepts

Network requests

A network request occurs when the user agent must use the network to service a single request.

If the user agent can service a request out of a local cache, that request MUST NOT result in a network request.

If the user agent follows redirects as part of a navigation, there MUST be separate network requests for each request in the redirect chain.

A request MUST NOT result in a network request if the user agent is known to be offline (i.e., when navigator.onLine returns false).

A request MUST NOT result in a network request if it is blocked due to mixed content or CORS failures. Any CORS-preflight request MUST result in its own network request.

For user agents that service requests according to the [[FETCH]] standard, a network request corresponds to one execution of the HTTP-network fetch algorithm.

Regardless of which fetch algorithm and which underlying application and transport protocols are used, servicing a network request consists of the following phases:

  1. DNS resolution: The user agent uses the Domain Name System [[RFC1034]] to resolve a domain name into an IP address of a server can that service HTTP requests to that domain.
  2. Secure connection establishment: The user agent opens a connection to the server, and establishes a secure channel over this connection.
  3. Transmission of request and response: Once the secure channel is established, the user agent can transmit the HTTP request, and receive the response from the server.

The only mandatory phase is the transmission of request and response; the other phases might not be needed for every network request. For instance, DNS results can be cached locally in the user agent, eliminating DNS resolution for future requests to the same domain. Similarly, HTTP persistent connections allow open connections to be shared for multiple requests to the same origin. However, if multiple phases occur, they will occur in the above order.

We would like to move the definition of these phases into [[FETCH]] so that they are more reusable.

A network request is successful if the user agent is able to receive a valid HTTP response from the server, and that response does not have a 4xx or 5xx status code.

A network request is failed if it is not successful.

Note that HTTP error responses (i.e., those with a 4xx or 5xx status code) are considered failures, so that they are subject to a NEL policy's failure sampling rate instead of its successful sampling rate.

Network errors

A network error is the error condition that caused a network request to fail.

Each network error has a type, which is a string.

Each network error has a phase, which describes which phase the error occurred in:

dns
the error occurred during DNS resolution
connection
the error occurred during secure connection establishment
application
the error occurred during the transmission of request and response

There are several predefined network error types defined in .

Network error reports

A network error report is a Reporting API report that describes a network error.

Network error reports have a report type of network-error.

Network error reports are NOT visible to ReportingObservers.

Network error reports are not visible to ReportingObservers because they are only intended to be visible to the administrator or owner of the server receiving the requests. If they were visible to ReportingObservers, then the reports would also be visible to the originator of the request. For cross-origin requests, this could leak information about the server's network configuration to parties outside of its control.

NEL policies

A NEL policy instructs a user agent whether to collect reports about network requests to an origin, and if so, where to send them. NEL policies are delivered to the user agent via HTTP response headers.

Each NEL policy has a received IP address, which is the IP address of the server that the user agent received this NEL policy from.

Each NEL policy has an origin.

Each NEL policy has a subdomains flag, which is either include or exclude.

Each NEL policy has a list of request headers and a list of response headers, each of which is a list of header names.

Each NEL policy has a reporting group, which is the name of the Reporting endpoint group that reports for this policy will be sent to.

Each NEL policy has a ttl representing the number of seconds the policy remains valid.

Each NEL policy has a creation which is the timestamp when the user agent received the policy.

A NEL policy is expired if its creation plus its ttl represents a time in the past.

Sampling rates

An origin that expects to serve a large volume of traffic might not be equipped to ingest NEL reports for every network request made to the origin. The origin can define sampling rates to limit the number of NEL reports that each user agent submits. Since successful requests should typically greatly outnumber failed requests, the origin can specify different sampling rates for each.

Each NEL policy has a successful sampling rate, which is a number between 0.0 and 1.0 inclusive.

Each NEL policy has a failure sampling rate, which is a number between 0.0 and 1.0 inclusive.

Policy cache

A conformant user agent MUST provide a policy cache, which is a storage mechanism that maintains a set of NEL policies, keyed by their origins.

This storage mechanism is opaque, vendor-specific, and not exposed to the web, but it MUST provide the following methods which will be used in the algorithms this document defines:

Policy delivery

A server MAY define a NEL policy for an origin it controls via the NEL HTTP response header.

NEL response header

The NEL response header is used to communicate an origin's NEL policy to the user agent. The ABNF (Augmented Backus-Naur Form) syntax for the NEL header is as follows:

NEL = json-field-value

The header's value is interpreted as an array of JSON objects, as defined by json-field-value. Each object in the array defines an NEL policy for the origin. The user agent MUST process the first valid policy in the array and ignore any additional policies in the array.

User agents MUST ignore any unknown or invalid field(s) or value(s) that do not conform to the syntax defined in this specification. A valid NEL header field MUST, at a minimum, contain one object with all of the "REQUIRED" fields defined in this specification.

The user agent MUST ignore the NEL header specified via a meta element to mitigate hijacking of error reporting via scripting attacks. The NEL policy MUST be delivered via the NEL response header.

The restriction on meta element is consistent with the [[CSP]] specification, which restricts reporting registration to HTTP header fields only for the same reasons.

The report_to member

The report_to member specifies the endpoint group that reports for this NEL policy will be sent to. The report_to member is REQUIRED to register a NEL policy, and OPTIONAL if the intent is to remove a previous registration – see max_age. If present, its value MUST be a string; any other type will result in a parse error.

To improve delivery of NEL reports, the server should set report_to to an endpoint group containing at least one endpoint in an alternative origin whose infrastructure is not coupled with the origin from which the resource is being fetched — otherwise network errors cannot be reported until the problem is solved, if ever — and provide multiple endpoints to provide alternatives if some endpoints are unreachable.

The max_age member

The REQUIRED max_age member specifies the lifetime of this NEL policy, as a non-negative integer number of seconds. Its value MUST be an non-negative integer; any other type will result in a parse error.

A value of 0 will cause any NEL policy for this origin to be removed from the policy cache.

To ensure delivery of NEL reports, the server should ensure that the Reporting endpoint group is also configured with a sufficiently high max_age. If the Reporting policy expires, NEL reports will not be delivered, even if the NEL policy has not expired.

The include_subdomains member

The OPTIONAL include_subdomains member is a boolean that enables this NEL policy for all subdomains of this origin. If no member named include_subdomains is present in the object, or its value is not true, the NEL policy will not be enabled for subdomains.

To ensure delivery of NEL reports for subdomains, the application should ensure that the Reporting endpoint group is also configured with include_subdomains enabled. If the Reporting policy is not, and there is not a separate Reporting policy for a given subdomain, NEL reports for that subdomain will not be delivered, even if the NEL policy includes the subdomain.

The success_fraction member

The OPTIONAL success_fraction member defines the sampling rate that should be applied to reports about successful network requests for this origin. If present, its value MUST be a number between 0.0 and 1.0, inclusive; any other value will result in a parse error. If this member is not present, the user agent will not collect NEL reports about successful network requests for this origin.

The failure_fraction member

The OPTIONAL failure_fraction member defines the sampling rate that should be applied to reports about failed network requests for this origin. If present, its value MUST be a number between 0.0 and 1.0, inclusive; any other value will result in a parse error. If this member is not present, the user agent will collect NEL reports about all failed network requests for this origin.

The request_headers member

The OPTIONAL request_headers member defines the list of request headers whose names and values will be included in network error reports about this origin. If present, its value MUST be a list of strings.

The response_headers member

The OPTIONAL response_headers member defines the list of response headers whose names and values will be included in network error reports about this origin. If present, its value MUST be a list of strings.

Process policy headers

Given a network request (request) and its corresponding response (response), this algorithm extracts a NEL policy for request's origin, and updates the policy cache accordingly.

  1. Abort these steps if any of the following conditions are true:
  2. Let origin be request's origin.
  3. Let header be the value of the response header whose name is NEL.
  4. Let list be the result of executing the algorithm defined in Section 4 of [[HTTP-JFV]] on header. If that algorithm results in an error, or if list is empty, abort these steps.
  5. Let item be the first element of list.
  6. If item has no member named max_age, or that member's value is not a number, abort these steps.
  7. If the value of item's max_age member is 0, then remove any NEL policy from the policy cache whose origin is origin, and skip the remaining steps.
  8. If item has no member named report_to, or that member's value is not a string, abort these steps.
  9. If item has a member named success_fraction, whose value is not a number in the range 0.0 to 1.0, inclusive, abort these steps.
  10. If item has a member named failure_fraction, whose value is not a number in the range 0.0 to 1.0, inclusive, abort these steps.
  11. If item has a member named request_headers, whose value is not a list, or if any element of that list is not a string, abort these steps.
  12. If item has a member named response_headers, whose value is not a list, or if any element of that list is not a string, abort these steps.
  13. Let policy be a new NEL policy whose properties are set as follows:

    received IP address
    the IP address of the server that the user agent received response from

    Plumb this through more explicitly in [[FETCH]].

    origin
    origin
    subdomains flag
    include if item has a member named include_subdomains whose value is true, exclude otherwise
    request headers
    the value of item's request_headers member
    response headers
    the value of item's response_headers member
    reporting group
    the value of item's report_to member
    ttl
    the value of item's max_age member
    creation
    the current timestamp
    successful sampling rate
    the value of item's success_fraction member, if present; 0.0 otherwise
    failure sampling rate
    the value of item's failure_fraction member, if present; 1.0 otherwise
  14. If there is already an entry in the policy cache for origin, replace it with policy; otherwise, insert policy into the policy cache for origin.

Report delivery

Choose a policy for an origin

Given an origin (origin), this algorithm determines which NEL policy in the policy cache should be used to generate reports for network requests to origin.

  1. If there is an entry in the policy cache for origin:
    1. Let policy be that entry.
    2. If policy is not expired, return it.
  2. For each parent origin that is a superdomain match of origin:
    1. If there is an entry in the policy cache for parent origin:
      1. Let policy be that entry.
      2. If policy is not expired, and its subdomains flag is include, return it.
  3. Return no policy.

Extract request headers

Given a network request (request) and a NEL policy (policy), this algorithm extracts header values from the request as instructed by the policy.

  1. Let headers be a new empty ECMAScript object.
  2. For each header name in policy's request headers list:
    1. If request's header list does not contain header name, skip to the next header name in the list.
    2. Let values be an empty ECMAScript list.
    3. For each header in request's header list whose name is header name, append header's value to values.
    4. Add a new property to headers whose name is header name and whose value is values.
  3. Return headers.

Extract response headers

Given a response (response) and a NEL policy (policy), this algorithm extracts header values from the response as instructed by the policy.

  1. Let headers be a new empty ECMAScript object.
  2. For each header name in policy's response headers list:
    1. If response's header list does not contain header name, AND response's trailer does not contain header name, skip to the next header name in the list.
    2. Let values be an empty ECMAScript list.
    3. For each header in response's header list whose name is header name, append header's value to values.
    4. For each header in response's trailer whose name is header name, append header's value to values.
    5. Add a new property to headers whose name is header name and whose value is values.
  3. Return headers.

Generate a network error report

Given a network request (request) and its corresponding response (response), this algorithm generates a report about request if instructed to by any matching NEL policy, and queues it for delivery.

  1. If the result of executing the "Is origin potentially trustworthy?" algorithm on request's origin is not Potentially Trustworthy, abort these steps.
  2. Let origin be request's origin.
  3. Let policy be the result of executing on origin. If policy is no policy, abort these steps.
  4. Determine the active sampling rate for this request:
  5. Decide whether or not to report on this request. Let roll be a random number between 0.0 and 1.0, inclusive. If rollsampling rate, abort these steps.
  6. Let report body be a new ECMAScript object with the following properties: [[ECMA-262]]
    referrer
    request's referrer, as determined by the referrer policy associated with its client.
    sampling_fraction
    sampling rate
    server_ip
    The IP address of the server to which the user agent sent the request, if available. Otherwise, an empty string.
    • A host identified by an IPv4 address is represented in dotted-decimal notation (a sequence of four decimal numbers in the range 0 to 255, separated by "."). [[RFC1123]]
    • A host identified by an IPv6 address is represented as an ordered list of eight 16-bit pieces (a sequence of `x:x:x:x:x:x:x:x`, where the 'x's are one to four hexadecimal digits of the eight 16-bit pieces of the address). [[RFC4291]]
    protocol
    The network protocol used to fetch the resource as identified by the ALPN Protocol ID, if available. Otherwise, "".
    method
    request's request method.
    request_headers
    The result of executing on request and policy.
    response_headers
    The result of executing on response and policy.
    status_code
    The status code of the HTTP response, if available. Otherwise, 0.
    elapsed_time
    The elapsed number of milliseconds between the start of the resource fetch and when it was completed or aborted by the user agent.
    phase
    If request failed, the phase of its network error. If request succeeded, "application".
    type
    If request failed, the type of its network error. If request succeeded, "ok".
  7. If report body's phase property is not dns, and report body's server_ip property is non-empty and not equal to policy's received IP address:
    1. Set report body's phase to dns.
    2. Set report body's type to dns.address_changed.
    3. Clear report body's request_headers, response_headers, status_code, and elapsed_time properties.
    4. Assert: All fields in report body that are derived from information not available during DNS resolution have been cleared.

    This step "downgrades" a NEL report if the IP addresses of the server and the policy don't match. This is a privacy protection, ensuring that NEL reports are only sent to the owner of the service that the report describes. If the IP addresses don't match, then the user agent can only verify that the NEL policy was sent by the owner of the origin's domain name; it cannot verify that the policy was sent by the owner of the server this domain name resolves to. We therefore downgrade the report to only contain information about DNS resolution. See and for more details.

  8. If origin is not equal to policy's origin, policy's subdomains flag is include, and report body's phase property is not dns, abort these steps.
  9. Queue the report for delivery via the Reporting API. [[!REPORTING]]

    type
    network-error
    data
    report body
    endpoint group
    policy's reporting group
    settings
    request's environment settings object
    url
    request's URL

Predefined network error types

There are several predefined network error types.

The user agent MAY extend this list with custom network error types — e.g. to accommodate new protocols, or more detailed error descriptions of existing ones. When doing so, the user agent SHOULD follow the dot-delimited pattern ([group].[optional-subgroup].[error-name]) for the type names to facilitate simple and consistent processing of the error reports — e.g. the collector may provide aggregation by category and/or one or multiple subgroups.

DNS resolution errors

All of the network errors in this section occur during DNS resolution, and therefore have a phase of dns.

dns.unreachable
DNS server is unreachable
dns.name_not_resolved
DNS server responded but is unable to resolve the address
dns.failed
Request to the DNS server failed due to reasons not covered by previous errors
dns.address_changed
Indicates that the resolved IP address for a request's origin has changed since the corresponding NEL policy was received

Secure connection establishment errors

All of the network errors in this section occur during secure connection establishment, and therefore have a phase of connection.

tcp.timed_out
TCP connection to the server timed out
tcp.closed
The TCP connection was closed by the server
tcp.reset
The TCP connection was reset
tcp.refused
The TCP connection was refused by the server
tcp.aborted
The TCP connection was aborted
tcp.address_invalid
The IP address is invalid
tcp.address_unreachable
The IP address is unreachable
tcp.failed
The TCP connection failed due to reasons not covered by previous errors
tls.version_or_cipher_mismatch
The TLS connection was aborted due to version or cipher mismatch
tls.bad_client_auth_cert
The TLS connection was aborted due to invalid client certificate
tls.cert.name_invalid
The TLS connection was aborted due to invalid name
tls.cert.date_invalid
The TLS connection was aborted due to invalid certificate date
tls.cert.authority_invalid
The TLS connection was aborted due to invalid issuing authority
tls.cert.invalid
The TLS connection was aborted due to invalid certificate
tls.cert.revoked
The TLS connection was aborted due to revoked server certificate
tls.cert.pinned_key_not_in_cert_chain
The TLS connection was aborted due to a key pinning error
tls.protocol.error
The TLS connection was aborted due to a TLS protocol error
tls.failed
The TLS connection failed due to reasons not covered by previous errors

Transmission of request and response errors

All of the network errors in this section occur during the transmission of request and response, and therefore have a phase of application.

http.error
The user agent successfully received a response, but it had a 4xx or 5xx status code
http.protocol.error
The connection was aborted due to an HTTP protocol error
http.response.invalid
Response is empty, has a content-length mismatch, has improper encoding, and/or other conditions that prevent user agent from processing the response
http.response.redirect_loop
The request was aborted due to a detected redirect loop
http.failed
The connection failed due to errors in HTTP protocol not covered by previous errors
abandoned
User aborted the resource fetch before it is complete
unknown
error type is unknown

Examples

Sample Policy Definitions

> GET / HTTP/1.1
> Host: example.com

< HTTP/1.1 200 OK
< ...
< Report-To: {"group": "network-errors", "max_age": 2592000,
              "endpoints": [{"url": "https://example.com/upload-reports"}]}
< NEL: {"report_to": "network-errors", "max_age": 2592000}
        

This NEL header defines a NEL policy, instructing the user agent to report network errors about example.com to the endpoint group named network-errors. The policy applies for 2592000 seconds (30 days).

Note that above registration will only succeed if the response is communicated from a potentially trustworthy origin.

> GET / HTTP/1.1
> Host: example.com

< HTTP/1.1 200 OK
< ...
< NEL: {"max_age": 0}
        

This NEL header instructs the user agent to remove any existing NEL policy for example.com.

Sample Network Error Reports

This section contains example network error reports the user agent might queue when a network error is encountered for an origin with a registered NEL policy. We show the full report payload that would be created by the [[!REPORTING]] API when uploading the report; the payload's body field contains the network error report body.

{
  "age": 0,
  "type": "network-error",
  "url": "https://www.example.com/",
  "body": {
    "sampling_fraction": 0.5,
    "referrer": "http://example.com/",
    "server_ip": "123.122.121.120",
    "protocol": "h2",
    "method": "GET",
    "request_headers": {},
    "response_headers": {},
    "status_code": 200,
    "elapsed_time": 823,
    "phase": "application",
    "type": "http.protocol.error"
  }
}
        

This report indicates that the user agent attempted to navigate from example.com to www.example.com, which successfully resolved to 123.122.121.120. However, while the user agent received a 200 response from the server via the HTTP/2 (h2) protocol, it encountered a protocol error in the exchange and was forced to abandon the navigation. The user agent aborted the navigation 823 milliseconds after it started. Finally, the user agent sent this report immediately after the network error was encountered – i.e. the report age is 0.

{
  "age": 0,
  "type": "network-error",
  "url": "https://widget.com/thing.js",
  "body": {
    "sampling_fraction": 1.0,
    "referrer": "https://www.example.com/",
    "server_ip": "",
    "protocol": "",
    "method": "GET",
    "request_headers": {},
    "response_headers": {},
    "status_code": 0,
    "elapsed_time": 143,
    "phase": "dns",
    "type": "dns.name_not_resolved"
  }
}
        

The above report indicates that the user agent attempted to fetch https://widget.com/thing.js from https://www.example.com/. However, the user agent was unable to resolve the DNS name (widget.com) and the request was aborted by the user agent after 143 milliseconds. Because a previous request to widget.com delivered a valid NEL policy, the user agent generates a network error report for this request. The report was uploaded immediately after the network error was encountered – i.e. the report age is 0.

DNS misconfiguration

> GET / HTTP/1.1
> Host: example.com

< HTTP/1.1 200 OK
< ...
< Report-To: {"group": "network-errors", "max_age": 2592000,
              "endpoints": [{"url": "https://example.com/upload-reports"}]}
< NEL: {"report_to": "network-errors", "max_age": 2592000, "include_subdomains": true}
        

This NEL header allows the owner of example.com to detect when they have misconfigured their DNS servers — for instance, when they have forgotten to add a new resource record resolving new-subdomain.example.com to an IP address. If a user agent tries to make a request to new-subdomain.example.com, it might generate the following report:

{
  "age": 0,
  "type": "network-error",
  "url": "https://new-subdomain.example.com/",
  "body": {
    "sampling_fraction": 1.0,
    "server_ip": "",
    "protocol": "http/1.1",
    "method": "GET",
    "request_headers": {},
    "response_headers": {},
    "status_code": 0,
    "elapsed_time": 48,
    "phase": "dns",
    "type": "dns.name_not_resolved"
  }
}
        

Monitoring cache validation

> GET / HTTP/1.1
> Host: example.com

< HTTP/1.1 200 OK
< ...
< Report-To: {"group": "network-errors", "max_age": 2592000,
              "endpoints": [{"url": "https://example.com/upload-reports"}]}
< NEL: {"report_to": "network-errors", "max_age": 2592000, "success_fraction": 1.0,
        "request_headers": ["If-None-Match"], "response_headers": ["ETag"]}
< ETag: 01234abcd
        

In this example, the owner of example.com uses ETag response headers to identify different versions of the resources hosted on the server. User agents can then use If-None-Match request headers to inform the server which version of a resource is presently cached client-side, allowing the server to avoid generating and sending the content of the resource if the client's existing copy is up to date.

By including request_headers and response_headers fields in the NEL header for this domain, the browser will include a copy of the If-None-Match request header and ETag response header in any NEL report that it creates for that request, allowing the site owner to track the effectiveness of their caching policies.

Given the above, consider the following sequence of events:

  1. The user agent sends a request to example.com, and receives a successful response from the server, with an ETag header indicating the version of the resource. The user agent will generate the following NEL report:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.1",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {},
        "response_headers": {
          "ETag": ["01234abcd"]
        },
        "status_code": 200,
        "elapsed_time": 1392,
        "phase": "application",
        "type": "ok"
      }
    }
                
  2. Some time later, the user agent sends another request to example.com. The user agent still has a copy of the original resource in its local cache, and includes its version in a If-None-Match request header. The server checks this version, notices that it is still current, and sends a 304 response informing the user agent that its cached copy of the resource is still valid. The user agent will generate the following report:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.1",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {
          "If-None-Match": ["01234abcd"]
        },
        "response_headers": {
          "ETag": ["01234abcd"]
        },
        "status_code": 304,
        "elapsed_time": 45,
        "phase": "application",
        "type": "ok"
      }
    }
                
  3. Even later, the user agent sends yet another request to example.com. The user agent still has the same copy of the resource in its local cache, and includes its version in a If-None-Match request header, as in the previous example. However, this time the server notices that there is a new version of the resource available. It generates the content of this resource, and sends it to the client, with the new version encoded in a new ETag response header value. The user agent will generate the following report:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.1",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {
          "If-None-Match": ["01234abcd"]
        },
        "response_headers": {
          "ETag": ["56789ef01"]
        },
        "status_code": 200,
        "elapsed_time": 935,
        "phase": "application",
        "type": "ok"
      }
    }
                

Origins with multiple IP addresses

For origins whose domain name resolves to multiple IP addresses, NEL will sometimes "downgrade" an error report, providing less information about the cause of the error, since it cannot verify that the owner of the origin is the same as the owner of the server handling the request.

As an example, assume that example.com is handled by three servers, each with a different IP address. The owner of the service configures DNS to resolve example.com to 192.0.2.1, 192.0.2.2, and 192.0.2.3, and relies on user agents to balance their requests across these three IP addresses. The service owner delivers the following NEL policy:

> GET / HTTP/1.1
> Host: example.com

< HTTP/1.1 200 OK
< ...
< Report-To: {"group": "network-errors", "max_age": 2592000,
              "endpoints": [{"url": "https://example.com/upload-reports"}]}
< NEL: {"report_to": "network-errors", "max_age": 2592000,
        "success_fraction": 1.0, "failure_fraction": 1.0}
        

Given the above, consider the following sequence of events:

  1. The user agent sends a request to 192.0.2.1, and receives a successful response from the server. This response includes the above NEL policy, and the user agent sets the policy's received IP address to 192.0.2.1. Since the received IP address matches the server's IP address (which it must for any successful request), it generates the following NEL report:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.1",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {},
        "response_headers": {},
        "status_code": 200,
        "elapsed_time": 57,
        "phase": "application",
        "type": "ok"
      }
    }
                
  2. The user agent sends a new request to 192.0.2.2, and receives another successful response. This response also includes the NEL policy, and the user agent updates the policy's received IP address to 192.0.2.2. Since the received IP address matches the server's IP address (which it must for any successful request), it generates the following NEL report:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.2",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {},
        "response_headers": {},
        "status_code": 200,
        "elapsed_time": 34,
        "phase": "application",
        "type": "ok"
      }
    }
                
  3. The user agent then tries to send a request to 192.0.2.3, but isn't able to establish a connection to the server. The user agent still has the NEL policy in the policy cache, and would ideally use this policy to generate a tcp.timed_out report about the failed network request. However, the because policy's received IP address (192.0.2.2) doesn't match the IP address that this request was sent to, the user agent cannot verify that the server at 192.0.2.3 is actually owned by the owners of example.com. The user agent must therefore downgrade the report to dns.address_changed:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.3",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {},
        "response_headers": {},
        "status_code": 0,
        "elapsed_time": 0,
        "phase": "dns",
        "type": "dns.address_changed"
      }
    }
                
  4. The user agent then tries to send another request to 192.0.2.1, but once again isn't able to establish a connection to the server. Even though the user agent received the NEL policy from 192.0.2.1 at some point in the past, the policy's received IP address only records where it was most recently received from — in this case, 192.0.2.2. The user agent must therefore downgrade the report to dns.address_changed:

    {
      "age": 0,
      "type": "network-error",
      "url": "https://example.com/",
      "body": {
        "sampling_fraction": 1.0,
        "server_ip": "192.0.2.1",
        "protocol": "http/1.1",
        "method": "GET",
        "request_headers": {},
        "response_headers": {},
        "status_code": 0,
        "elapsed_time": 0,
        "phase": "dns",
        "type": "dns.address_changed"
      }
    }
                

Use cases

Reporting of Navigation Failures

A navigation request initiated by the user (e.g. via a click on a link, direct input via the location bar, script-initiated due to user interaction, etc.) may fail due any number of connectivity reasons: DNS failure, TCP error, TLS protocol violation, and so on. These errors may be caused by network misconfiguration, transient routing issues, server downtime, malware or other attacks against the user, etc.

In such cases the destination host is often left unaware of the failed navigation since, by definition, it cannot see the request reach its infrastructure and it is unable to investigate the problem. To address this, the host can register an NEL policy with the user agent, which specifies where reports of such failures should be delivered such that they can be investigated.

Reporting of First-party Subresource Fetch Failures

A typical application requires dozens of resources, the fetching of which is typically initiated via HTML, CSS, or JavaScript. The application requesting such resources can observe failures of most such fetches (e.g. via `onerror` callbacks), but it does not have access to the detailed network error report of why the failure has occurred - e.g. DNS failure, TCP error, TLS protocol violation, etc.

To address this, the application can register relevant NEL policies with the user agent for the first-party hosts from which the subresources are being fetched. Then, if such a policy is present and a network error is encountered for a resource from an origin with a registered NEL policy, the user agent will report the detailed network error report and enable the application developers to investigate the error.

Reporting of Third-party Subresource Fetch Failures

In the case where a resource is embedded by a third party, the provider of the resource is often unable to instrument and observe the failure. For example, if `example.com` embeds a `widget.com/thing.js` resource on its site, and the user visiting `example.com` fails to fetch such resource due to a network error, the `widget.com` host is both unaware of the failure and unable to detect it.

To address this, `widget.com` can register an NEL policy for its host. Then, if such policy is present and a network error is encountered while fetching a resource — regardless of whether it is being requested from a first-party or third-party origin — from the origin with a registered NEL policy, the user agent will report the network error and enable the provider to investigate the error.

Privacy Considerations

NEL provides network error reports that could expose new information about the user's network configuration. For example, an attacker could abuse NEL reporting to probe the user's network configuration, or to scan for servers on the user's internal network. Also, similar to HSTS, HPKP, and pinned CSP policies, the stored NEL policy could be used as a "supercookie" by setting a distinct policy with a custom (per-user) reporting URI to act as an identifier in combination with (or instead of) HTTP cookies.

To mitigate some of the above risks, NEL registration is restricted to potentially trustworthy origins, and delivery of network error reports is similarly restricted to potentially trustworthy origins. This disallows a transient HTTP MITM from trivially abusing NEL as a persistent tracker.

NEL is intended to augment existing server-side monitoring. NEL reports should only be sent to the owner of the service being requested. For errors that occur during DNS resolution, NEL reports are only generated when the NEL policy was received from the owner of the domain namespace tree that contains the policy origin. For errors that occur during secure connection establishment or transmission of request and response, NEL reports are only generated when the NEL policy was received from the owner of the server that the request was sent to.

This rationale explains the treatment of the received IP address and subdomains flag of a NEL policy. By checking that the policy's received IP address matches the IP address of the server, NEL extends the trust boundary of the policy to include not just the policy's origin, but also the specific server that the user agent is communicating with. This helps prevent (for instance) DNS rebinding attacks, where an attacker delivers a long-lived NEL policy from a server that they own, and then changes their name servers to resolve the policy origin to a server they don't control. Without the received IP address verification, this would cause user agents to send reports about the second server to the attacker.

Similarly, subdomain NEL policies are limited, and can only be used to generate reports about subdomains of the policy origin during the DNS resolution phase of a request. During this phase, there is no server to verify ownership of, and the fact that the policy was received from a superdomain of the request's origin is enough to establish ownership of the error. This allows the owners of a particular portion of the domain namespace tree to use NEL to detect errors, while preventing them from using malicious DNS entries to collect information about servers they don't control.

To prevent information leakage, NEL reports about a request do not contain any information that is not visible to the server when processing the request. For errors during DNS resolution, a NEL report only contains information available from DNS itself. This prevents servers from abusing NEL to collect more information about their users than they already have access to.

As an example, NEL reports specifically do not contain any information about which DNS resolver was used to resolve a request's domain name into an IP address.

In addition to above restrictions, the user agents MUST:

When deploying NEL the developer SHOULD consider privacy implications of NEL reports delivered to the specified collectors. For example, reports may contain URLs with sensitive data (e.g. "Capability URLs") that may need special precautions (see [[!CAPABILITY-URLS]]), and may require the developer to operate their own NEL collectors to prevent reporting of such URLs to third parties.

IANA Considerations

The permanent message header field registry should be updated with the following registrations ([[RFC3864]]):

NEL

Header field name
NEL
Applicable protocol
http
Status
standard
Author/Change controller
W3C
Specification document
This specification (see NEL response header)

Acknowledgments

This document reuses text from the [[CSP]] and [[RFC6797]] specification, as permitted by the licenses of those specifications. Additionally, sincere thanks to Julia Tuttle, Chris Bentzel, Todd Reifsteck, Aaron Heady, and Mark Nottingham for their helpful comments and contributions to this work.