| Internet-Draft | DNS-ER | July 2022 |
| Arends & Larson | Expires 9 January 2023 | [Page] |
DNS Error Reporting is a lightweight error reporting mechanism that provides the operator of an authoritative server with reports on DNS resource records that fail to resolve or validate, that a Domain Owner or DNS Hosting organization can use to improve domain hosting. The reports are based on Extended DNS Errors [RFC8914].¶
When a domain name fails to resolve or validate due to a misconfiguration or an attack, the operator of the authoritative server may be unaware of this. To mitigate this lack of feedback, this document describes a method for a validating recursive resolver to automatically signal an error to an agent specified by the authoritative server. DNS Error Reporting uses the DNS to report errors.¶
Another lack of feedback occurs when validation was successful, or when there is no error to report. This positive feedback may be helpful to show that a deployment was successful. This document introcudes an extended DNS error "NO ERROR".¶
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 9 January 2023.¶
Copyright (c) 2022 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.¶
When an authoritative server serves a stale DNSSEC signed zone, the cryptographic signatures over the resource record sets (RRsets) may have lapsed. A validating recursive resolver will fail to validate these resource records.¶
Similarly, when there is a mismatch between the DS records at a parent zone and the key signing key at the child zone, a validating recursive resolver will fail to authenticate records in the child zone.¶
These are two of several failure scenarios that may go unnoticed for some time by the operator of a zone.¶
There is no direct relationship between operators of validating recursive resolvers and authoritative servers. Outages are often noticed indirectly, by end users, and reported via social media, if reported at all.¶
When records fail to validate there is no facility to report this failure in an automated way. If there is any indication that an error or warning has happened, it is buried in log files of the validating resolver, if these errors are logged at all.¶
This document describes a facility that can be used by validating recursive resolvers to report errors in an automated way. In addition, successful validation, or a lack of errors can also be reported in an automated way.¶
It allows an authoritative server to signal a reporting agent where the validating recursive resolver can report issues if it is configured to do so. The signal also indicates that the reporting agent is interested in successful validation, or lack of errors.¶
The burden of reporting a failure falls on the validating recursive resolver. It is important that the effort needed to report failure is low, with minimal impact to its main functions. To accomplish this goal, the DNS itself is utilized to report the error.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].¶
Reporting Resolver: In the context of this document, the term reporting resolver is used as a shorthand for a validating recursive resolver that supports DNS Error Reporting.¶
Reporting Query: The DNS query used to report an error is called a reporting query. A reporting query is for DNS resource record type NULL. The details of the error report are encoded in the QNAME of the reporting query.¶
Reporting Agent: A facility responsible for receiving error reports on behalf of authoritative servers. This facility is indicated by a domain name.¶
Reporting Agent Domain: a domain name which the reporting resolver includes in the QNAME of the reporting query.¶
Positive Feedback: A report that indicates that no error occurred.¶
An authoritative server indicates support for DNS Error Reporting by including an EDNS0 option with OPTION-CODE [TBD] [RFC Editor: change TBD to the proper code when assigned by IANA.], a flag to indicate a request for positive feedback and the REPORTING AGENT DOMAIN in DNS wireformat in the option's payload. The authoritative server MUST NOT include this option in the response if the configured reporting agent domain is empty or the null label (the root).¶
The positive feedback flag indicates that the reporting agent wants to receive extended DNS error [TBD] that indicates that no error occurred. This extended DNS error is defined in this document.¶
When a reporting resolver sends a reporting query to report an error, it MUST NOT include the EDNS0 Error Reporting option in the reporting query. This avoids additional compounding error reporting when there is an issue with the reporting agent domain.¶
To report an error, the reporting resolver encodes the error report in the QNAME of the reporting query. The reporting resolver builds this QNAME by concatenating the _er label, the extended error code [RFC8914], the QTYPE and the QNAME that resulted in failure, the label "_er" again, and the reporting agent domain. See the example in section 4.2.¶
The resulting concatenated domain name is sent as a standard DNS query for DNS resource record type NULL by the reporting resolver. This query MUST NOT have the EDNS0 option code [TBD] set to avoid compounding error notifications.¶
The query will ultimately arrive at the authoritative server for the reporting agent domain. A NODATA negative response is returned by the authoritative server of the reporting agent domain, which in turn can be cached by the reporting resolver.¶
This caching is essential. It ensures that the number of reports sent by a reporting resolver for the same problem is dampened, i.e. once per TTL, however, certain optimizations such as [RFC8020] and [RFC8198] may reduce the number of error reporting queries as well.¶
The reporting resolver may utilize various caching optimizations that inhibit subsequent error reporting to the same reporting agent domain.¶
If the authoritative server for the reporting agent domain were to respond with NXDOMAIN (name error), [RFC8020] rules state that any name at or below that domain should be considered unreachable, and negative caching would prohibit subsequent queries for anything at or below that domain for a period of time, depending on the negative TTL [RFC2308].¶
Since the authoritative server for an agent domain may not know the contents of all the zones it acts as an agent for, it is essential that the authoritative server does not respond with NXDOMAIN, as that may inhibit subsequent queries. The use of a wildcard domain name [RFC4592] in the zone for the agent domain will ensure the RCODE is consistently NOERROR.¶
Considering the Resource Record type for this wildcard record, type NULL is prohibited in master zone files [RFC1035]. However, any type that is not special according to [RFC4592] section 4 will do, such as a TXT record with an email address for the reporting agent in the RDATA.¶
Wildcard expansion occurs, even if the QTYPE is not for the type owned by the wildcard domain name. The response is a "no error, but no data" response ([RFC4592], section 2.2.1.) that contains a NOERROR RCODE and empty answer section. Note that reporting resolvers are not expected to query for this TXT record, since reporting queries use type NULL. This record is solely present to ensure a NODATA response is returned in response to reporting queries.¶
When the zone for the reporting agent domain is signed, a resolver may utilize aggressive negative caching, discussed in [RFC8198]. This optimization makes use of NSEC and NSEC3 (without opt-out) records and allows the resolver to do the wildcard synthesis. When this happens, the resolver may not send subsequent queries as it will be able to synthesize a response from previously cached material.¶
A solution is to avoid DNSSEC for the reporting agent domain. Signing the agent domain will incur an additional burden on the reporting resolver, as it has to validate the response. However, this response has no utility to the reporting resolver.¶
The domain broken.test is hosted on a set of authoritative servers. One of these serves a stale version. This authoritative server has a severity level of 1 and a reporting agent configured: a01.reporting-agent.example.¶
The reporting resolver is unable to validate the broken.test RRSet for type A, due to an RRSIG record with an expired signature.¶
The reporting resolver constructs the QNAME _er.7.1.broken.test._er.a01.reporting-agent.example and resolves it. This QNAME indicates extended DNS error 7 occurred while trying to validate broken.test type 1 (A) record.¶
After this query is received at one of the authoritative servers for the reporting agent domain (a01.reporting-agent.example), the reporting agent (the operators of the authoritative server for a01.reporting-agent.example) determines that the authoritative server for the broken.test zone suffers from an expired signature record (extended error 7) for type A for the domain name broken.test. The reporting agent can contact the operators of broken.test to fix the issue.¶
This method uses an EDNS0 [RFC6891] option to indicate support for sending DNS error reports and responding with the Reporting Agent Domain in DNS messages. The option is structured as follows:¶
1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OPTION-CODE = TBD | OPTION-LENGTH |
+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
|F| REPORTING AGENT DOMAIN /
+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
¶
Field definition details:¶
The various errors that a reporting resolver may encounter are listed in [RFC8914]. Note that not all listed errors may be supported by the reporting resolver. This document does not specify what is an error and what is not.¶
The DNS class is not specified in the error report.¶
Reporting Resolvers may have a configuration that allows the following:¶
The reporting resolver MUST NOT use DNS error reporting to report a failure in resolving the reporting query.¶
The reporting resolver MUST NOT use DNS error reporting if the authoritative server has an empty Reporting Agent Domain field in the EDNS Error Reporting option.¶
The reporting resolver should limit the amount of positive feedback send.¶
The QNAME for the reporting query is constructed by concatenating the following elements, appending each successive element in the list to the right-hand side of the QNAME:¶
If the resulting reporting query QNAME would exceed 255 octets, it MUST NOT be sent.¶
The "_er" labels allow the reporting agent to quickly differentiate between the agent domain and the faulty query name. When the specified agent domain is empty, or a NULL label (despite being not allowed in this specification), the reporting query will have "_er" as a top-level domain as a result and not the original query. Lastly, the purpose of the first "_er" label is to indicate that a complete reporting query has been received, instead of a shorter reporting query due to query minimization.¶
It is RECOMMENDED that the reporting agent zone uses a wildcard DNS record of type TXT with an arbitrary string in the RDATA and a TTL of at least one hour.¶
It is RECOMMENDED that the reporting agent domain be kept relatively short to allow for a longer QNAME in the reporting query.¶
IANA is requested to assign the following DNS EDNS0 option code registry:¶
Value Name Status Reference
----- ---------------- -------- ---------------
TBD DNS ERROR REPORT Standard [this document]
¶
[RFC Editor: change TBD to the proper code when assigned by IANA.]¶
IANA is requested to assign the following Underscored and Globally Scoped DNS Node Name registry:¶
RR Type _NODE NAME Reference
----- ---------- ---------
TXT _er [this document]
¶
IANA is requested to assign the following value the "Extended DNS Error Codes" registry:¶
INFO-CODE: [TBD]¶
Purpose: No Error¶
Reference: This document¶
Use of DNS Error Reporting may expose local configuration mistakes in the reporting resolver, such as stale DNSSEC trust anchors to the reporting agent.¶
DNS Error reporting SHOULD be done using DNS Query Name Minimization [RFC7816] to improve privacy.¶
DNS Error Reporting is done without any authentication between the reporting resolver and the authoritative server of the agent domain. Authentication significantly increases the burden on the reporting resolver without any benefit to the reporting agent, authoritative server or reporting resolver.¶
The method described in this document will cause additional queries by the reporting resolver to authoritative servers in order to resolve the reporting query.¶
This method can be abused by deploying broken zones with agent domains that are delegated to servers operated by the intended victim in combination with open resolvers [RFC8499].¶
This document is based on an idea by Roy Arends and David Conrad. The authors would like to thank Peter van Dijk, Vladimir Cunat, Paul Hoffman, Libor Peltan, Matthijs Mekking, Willem Toorop, Tom Carpay, Dick Franks, Benno Overeinder and Petr Spacek for their contributions.¶