Problem Details for HTTP APIs

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶

This Internet-Draft will expire on 30 July 2023.¶

Copyright Notice

Copyright (c) 2023 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

HTTP status codes (Section 15 of [HTTP]) cannot always convey enough information about errors to be helpful. While humans using Web browsers can often understand an HTML [HTML5] response content, non-human consumers of HTTP APIs have difficulty doing so.¶

To address that shortcoming, this specification defines simple JSON [JSON] and XML [XML] document formats to describe the specifics of problem(s) encountered -- "problem details".¶

For example, consider a response indicating that the client's account doesn't have enough credit. The API's designer might decide to use the 403 Forbidden status code to inform HTTP-generic software (such as client libraries, caches, and proxies) of the response's general semantics. API-specific problem details (such as why the server refused the request and the applicable account balance) can be carried in the response content, so that the client can act upon them appropriately (for example, triggering a transfer of more credit into the account).¶

This specification identifies the specific "problem type" (e.g., "out of credit") with a URI [URI]. HTTP APIs can use URIs under their control to identify problems specific to them, or can reuse existing ones to facilitate interoperability and leverage common semantics (see Section 4.2).¶

Problem details can contain other information, such as a URI identifying the problem's specific occurrence (effectively giving an identifier to the concept "The time Joe didn't have enough credit last Thursday"), which can be useful for support or forensic purposes.¶

The data model for problem details is a JSON [JSON] object; when serialized as a JSON document, it uses the "application/problem+json" media type. Appendix B defines an equivalent XML format, which uses the "application/problem+xml" media type.¶

When they are conveyed in an HTTP response, the contents of problem details can be negotiated using proactive negotiation; see Section 12.1 of [HTTP]. In particular, the language used for human-readable strings (such as those in title and description) can be negotiated using the Accept-Language request header field (Section 12.5.4 of [HTTP]), although that negotiation may still result in a non-preferred, default representation being returned.¶

Note that problem details are (naturally) not the only way to convey the details of a problem in HTTP. If the response is still a representation of a resource, for example, it's often preferable to describe the relevant details in that application's format. Likewise, defined HTTP status codes cover many situations with no need to convey extra detail.¶

This specification's aim is to define common error formats for applications that need one so that they aren't required to define their own, or worse, tempted to redefine the semantics of existing HTTP status codes. Even if an application chooses not to use it to convey errors, reviewing its design can help guide the design decisions faced when conveying errors in an existing format.¶

See Appendix D for a list of changes from RFC 7807.¶

2. Notational Conventions

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

3. The Problem Details JSON Object

The canonical model for problem details is a JSON [JSON] object. When serialized in a JSON document, that format is identified with the "application/problem+json" media type.¶

For example:¶

POST /purchase HTTP/1.1
Host: store.example.com
Content-Type: application/json
Accept: application/json, application/problem+json

{
  "item": 123456,
  "quantity": 2
}

HTTP/1.1 403 Forbidden
Content-Type: application/problem+json
Content-Language: en

{
 "type": "https://example.com/probs/out-of-credit",
 "title": "You do not have enough credit.",
 "detail": "Your current balance is 30, but that costs 50.",
 "instance": "/account/12345/msgs/abc",
 "balance": 30,
 "accounts": ["/account/12345",
              "/account/67890"]
}

Here, the out-of-credit problem (identified by its type) indicates the reason for the 403 in "title", identifies the specific problem occurrence with "instance", gives occurrence-specific details in "detail", and adds two extensions: "balance" conveys the account's balance, and "accounts" lists links where the account can be topped up.¶

When designed to accommodate it, problem-specific extensions can convey more than one instance of the same problem type. For example:¶

POST /details HTTP/1.1
Host: account.example.com
Accept: application/json

{
  "age": 42.3,
  "profile": {
    "color": "yellow"
  }
}

HTTP/1.1 400 Bad Request
Content-Type: application/problem+json
Content-Language: en

{
 "type": "https://example.net/validation-error",
 "title": "Your request is not valid.",
 "errors": [
             {
               "detail": "must be a positive integer",
               "pointer": "#/age"
             },
             {
               "detail": "must be 'green', 'red' or 'blue'",
               "pointer": "#/profile/color"
             }
          ]
  }

The fictional problem type here defines the "errors" extension, an array that describes the details of each validation error. Each member is an object containing "detail" to describe the issue, and "pointer" to locate the problem within the request's content using a JSON Pointer [JSON-POINTER].¶

When an API encounters multiple problems that do not share the same type, it is RECOMMENDED that the most relevant or urgent problem be represented in the response. While it is possible to create generic "batch" problem types that convey multiple, disparate types, they do not map well into HTTP semantics.¶

Note also that the API has responded with the application/problem+json type, even though the client did not list it in Accept, as is allowed by HTTP (see Section 12.5.1 of [HTTP]).¶

tag:mnot@mnot.net,2021-09-17:OutOfLuck

4. Defining New Problem Types

When an HTTP API needs to define a response that indicates an error condition, it might be appropriate to do so by defining a new problem type.¶

Before doing so, it's important to understand what they are good for, and what's better left to other mechanisms.¶

Problem details are not a debugging tool for the underlying implementation; rather, they are a way to expose greater detail about the HTTP interface itself. Designers of new problem types need to carefully take into account the Security Considerations (Section 5), in particular, the risk of exposing attack vectors by exposing implementation internals through error messages.¶

Likewise, truly generic problems -- i.e., conditions that might apply to any resource on the Web -- are usually better expressed as plain status codes. For example, a "write access disallowed" problem is probably unnecessary, since a 403 Forbidden status code in response to a PUT request is self-explanatory.¶

Finally, an application might have a more appropriate way to carry an error in a format that it already defines. Problem details are intended to avoid the necessity of establishing new "fault" or "error" document formats, not to replace existing domain-specific formats.¶

That said, it is possible to add support for problem details to existing HTTP APIs using HTTP content negotiation (e.g., using the Accept request header to indicate a preference for this format; see [HTTP], Section 12.5.1).¶

New problem type definitions MUST document:¶

1. a type URI (typically, with the "http" or "https" scheme),¶
2. a title that appropriately describes it (think short), and¶
3. the HTTP status code for it to be used with.¶

Problem type definitions MAY specify the use of the Retry-After response header ([HTTP], Section 10.2.3) in appropriate circumstances.¶

A problem's type URI SHOULD resolve to HTML [HTML5] documentation that explains how to resolve the problem.¶

A problem type definition MAY specify additional members on the problem details object. For example, an extension might use typed links [WEB-LINKING] to another resource that machines can use to resolve the problem.¶

If such additional members are defined, their names SHOULD start with a letter (ALPHA, as per [ABNF], Appendix B.1) and SHOULD comprise characters from ALPHA, DIGIT ([ABNF], Appendix B.1), and "_" (so that it can be serialized in formats other than JSON), and they SHOULD be three characters or longer.¶

5. Security Considerations

When defining a new problem type, the information included must be carefully vetted. Likewise, when actually generating a problem -- however it is serialized -- the details given must also be scrutinized.¶

Risks include leaking information that can be exploited to compromise the system, access to the system, or the privacy of users of the system.¶

Generators providing links to occurrence information are encouraged to avoid making implementation details such as a stack dump available through the HTTP interface, since this can expose sensitive details of the server implementation, its data, and so on.¶

The "status" member duplicates the information available in the HTTP status code itself, bringing the possibility of disagreement between the two. Their relative precedence is not clear, since a disagreement might indicate that (for example) an intermediary has changed the HTTP status code in transit (e.g., by a proxy or cache). Generic HTTP software (such as proxies, load balancers, firewalls, and virus scanners) are unlikely to know of or respect the status code conveyed in this member.¶

6. IANA Considerations

Please update the "application/problem+json" and "application/problem+xml" registrations in the "Media Types" registry to refer to this document.¶

Please create the "HTTP Problem Types" registry as specified in Section 4.2, and populate it with "about:blank" as per Section 4.2.1.¶

Appendix A. JSON Schema for HTTP Problems

This section presents a non-normative JSON Schema [JSON-SCHEMA] for HTTP Problem Details. If there is any disagreement between it and the text of the specification, the latter prevails.¶

# NOTE: '\' line wrapping per RFC 8792
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "A problem object RFC 7807bis",
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "format": "uri-reference",
      "description": "A URI reference RFC3986 that identifies the \
problem type."
    },
    "title": {
      "type": "string",
      "description": "A short, human-readable summary of the \
problem type. It SHOULD NOT change from occurrence to occurrence \
of the problem, except for purposes of localization (e.g., using \
proactive content negotiation; see RFC7231, Section 3.4)"
    },
    "status": {
      "type": "integer",
      "description": "The HTTP status code (RFC7231, Section 6) \
generated by the origin server for this occurrence of the problem.",
      "minimum": 100,
      "maximum": 599
    },
    "detail": {
      "type": "string",
      "description": "A human-readable explanation specific to \
this occurrence of the problem."
    },
    "instance": {
      "type": "string",
      "format": "uri-reference",
      "description": "A URI reference that identifies the \
specific occurrence of the problem. It may or may not yield \
further information if dereferenced."
    }
  }
}

Appendix B. HTTP Problems and XML

HTTP-based APIs that use XML [XML] can express problem details using the format defined in this appendix.¶

The RELAX NG schema [ISO-19757-2] for the XML format is:¶

   default namespace ns = "urn:ietf:rfc:7807"

   start = problem

   problem =
     element problem {
       (  element  type            { xsd:anyURI }?
        & element  title           { xsd:string }?
        & element  detail          { xsd:string }?
        & element  status          { xsd:positiveInteger }?
        & element  instance        { xsd:anyURI }? ),
       anyNsElement
     }

   anyNsElement =
     (  element    ns:*  { anyNsElement | text }
      | attribute  *     { text })*

Note that this schema is only intended as documentation, and not as a normative schema that captures all constraints of the XML format. It is possible to use other XML schema languages to define a similar set of constraints (depending on the features of the chosen schema language).¶

The media type for this format is "application/problem+xml".¶

Extension arrays and objects are serialized into the XML format by considering an element containing a child or children to represent an object, except for elements that contain only child element(s) named 'i', which are considered arrays. For example, the example above appears in XML as follows:¶

HTTP/1.1 403 Forbidden
Content-Type: application/problem+xml
Content-Language: en

<?xml version="1.0" encoding="UTF-8"?>
<problem xmlns="urn:ietf:rfc:7807">
  <type>https://example.com/probs/out-of-credit</type>
  <title>You do not have enough credit.</title>
  <detail>Your current balance is 30, but that costs 50.</detail>
  <instance>https://example.net/account/12345/msgs/abc</instance>
  <balance>30</balance>
  <accounts>
    <i>https://example.net/account/12345</i>
    <i>https://example.net/account/67890</i>
  </accounts>
</problem>

This format uses an XML namespace, primarily to allow embedding it into other XML-based formats; it does not imply that it can or should be extended with elements or attributes in other namespaces. The RELAX NG schema explicitly only allows elements from the one namespace used in the XML format. Any extension arrays and objects MUST be serialized into XML markup using only that namespace.¶

When using the XML format, it is possible to embed an XML processing instruction in the XML that instructs clients to transform the XML, using the referenced XSLT code [XSLT]. If this code is transforming the XML into (X)HTML, then it is possible to serve the XML format, and yet have clients capable of performing the transformation display human-friendly (X)HTML that is rendered and displayed at the client. Note that when using this method, it is advisable to use XSLT 1.0 in order to maximize the number of clients capable of executing the XSLT code.¶

Appendix C. Using Problem Details with Other Formats

In some situations, it can be advantageous to embed problem details in formats other than those described here. For example, an API that uses HTML [HTML5] might want to also use HTML for expressing its problem details.¶

Problem details can be embedded in other formats either by encapsulating one of the existing serializations (JSON or XML) into that format or by translating the model of a problem detail (as specified in Section 3) into the format's conventions.¶

For example, in HTML, a problem could be embedded by encapsulating JSON in a script tag:¶

<script type="application/problem+json">
  {
   "type": "https://example.com/probs/out-of-credit",
   "title": "You do not have enough credit.",
   "detail": "Your current balance is 30, but that costs 50.",
   "instance": "/account/12345/msgs/abc",
   "balance": 30,
   "accounts": ["/account/12345",
                "/account/67890"]
  }
</script>

or by inventing a mapping into RDFa [RDFA].¶

This specification does not make specific recommendations regarding embedding problem details in other formats; the appropriate way to embed them depends both upon the format in use and application of that format.¶

Appendix D. Changes from RFC 7807

This revision has made the following changes:¶

• Section 4.2 introduces a registry of common problem type URIs¶
• Section 3 clarifies how multiple problems should be treated¶
• Section 3.2 reserves a prefix for future standards-defined object members¶
• Section 3.1.1 provides guidance for using type URIs that cannot be dereferenced¶

Acknowledgements

The authors would like to thank Jan Algermissen, Subbu Allamaraju, Mike Amundsen, Roy Fielding, Eran Hammer, Sam Johnston, Mike McCall, Julian Reschke, and James Snell for review of this specification.¶

Authors' Addresses