Locate a timing question without claiming a measured queue
Received Headers and Email Delivery Delays
A message that appears late creates a tempting question: which server held it? Headers can help narrow the investigation, but they do not supply a stopwatch for every queue. Received fields are transport declarations with timestamps, not independently observed events. Read their order, clock information, and gaps together before assigning a cause. A useful report identifies the adjacent declarations worth checking and names what their text cannot establish.
Use the Email Header Analyzer with raw header text. Load example fills the input without running it; select Analyze headers to inspect the declarations. Clear removes the working result. No connection is made to the named mail servers, so this is an offline reading exercise, not a live route test.
Reverse the source order without sorting away the evidence
SMTP receivers normally prepend their Received field, placing the newest added trace above the existing ones. RFC 5321 describes that insertion behavior, while RFC 5322 defines the message trace format. To present the apparent progression, the analyzer reverses the Received fields from the supplied source: the last occurrence is displayed first. Oldest declaration first describes that reversal, not a guarantee about true chronology. The software does not sort by date. Preserving this distinction matters because a strange timestamp is part of the evidence you came to examine. Reordering rows until their clocks look tidy would conceal the very discrepancy that may need investigation.
Consider three source occurrences from top to bottom, labelled A, B, and C. The route presentation is C, B, A even if their written dates suggest a different order. Those labels are illustrative, not extra header fields. Compare C with B, then B with A, while retaining the original occurrences for reference. If B has a time earlier than C, do not drag it elsewhere or call the reversal a parser error. The source and timestamps disagree about apparent progression. An administrator can investigate that disagreement only if the report preserves both kinds of information rather than silently choosing the more convenient one.
Within a declaration, from and by identify claimed participants, while with, id, and for may provide protocol, correlation, or recipient context. A bounded token summary is not a verified network map. Missing tokens stay missing, and a hostname is not looked up to prove ownership, location, or reachability. Names supplied by a peer and details added by a receiving system are not equally established merely because they share a line. Keep unfamiliar syntax beside the raw declaration instead of filling gaps with a likely provider. A URL question is separate: Link Inspector can inspect a deliberately supplied nonsecret URL, not certify this route.
Read the timestamp and its offset before comparing clocks
The transport timestamp follows the Received field delimiter semicolon; a date-like string inside other descriptive text is not automatically that timestamp. Folded lines still belong to their original field, so visual line breaks do not create extra hops. The parser interprets supported date syntax rather than treating every digit sequence as a clock. Start your comparison with the complete declared date, including its calendar day and offset. A local time of 00:03 can follow 23:58 on the previous day without any inconsistency. Ignoring the date while comparing hour-and-minute displays is an easy way to manufacture an overnight delay or an apparent backwards transition.
An explicit numeric offset permits a common time reference when the date is otherwise usable. For example, 10:05 at +0200 corresponds to 08:05 at +0000. A following declaration at 08:07 +0000 is two minutes later, not almost two hours earlier. This is arithmetic on declared values, not confirmation that either server clock was accurate. Keep the raw offset in your notes even when the display also offers an interpreted value. Do not replace it with your own current browser zone, infer a server zone from its domain ending, or apply present-day daylight-saving assumptions to a timestamp that already supplies an explicit offset.
Unknown time information must remain unknown. A missing or unparseable timestamp does not represent zero, the current time, or an invitation to borrow a neighboring value. This analyzer treats -0000 as unknown timezone information and does not silently relabel it as an ordinary explicit UTC offset. That preserves the distinction from +0000 for this troubleshooting view. If an adjacent declaration lacks a usable timestamp or offset, its interval cannot be established here. Do not edit an unknown zone to a guessed value merely to obtain a complete-looking timeline. Ask for the original source or relevant server logs and leave the uncertainty visible until evidence resolves it.
Label adjacent signed differences as declaration arithmetic
For adjacent entries in reversed source order, the signed difference is the next declaration timestamp minus the preceding declaration timestamp, where both are usable. Suppose the first says 09:02:10 +0000 and the next says 09:17:40 +0000 on the same day. Their declared timestamp difference is +15 minutes 30 seconds. That is the correct label for the arithmetic. Calling it fifteen minutes in a particular queue would add a causal claim the headers do not measure. The two stamps come from separate declarations and can bracket processing, retries, transport, clock error, or other activity not itemized in the supplied block.
A negative difference is especially informative because it flags a clock or ordering question instead of a negative delivery duration. If the next declaration says 09:01:30 while the preceding one says 09:02:10 with the same known offset, the difference is -40 seconds. Preserve the sign. Possible questions include whether clocks disagreed, the source was rearranged, or a declaration was malformed or invented; the parser cannot choose the cause. Do not convert the result to an absolute value, clamp it to zero, or date-sort it out of sight. Each of those edits would replace the observed textual inconsistency with a cleaner but misleading account.
A zero difference does not prove an instantaneous handoff. Two systems can declare the same second while their actual event times differ within that precision, and the declarations need not come from synchronized clocks. Equally, several positive differences do not justify a measured end-to-end delivery total. A missing middle timestamp breaks the adjacent comparison; the tool must not bridge across it and pretend the intervening segment was timed. You may describe a span between two usable declarations separately, but identify both endpoints and the unknown interior. Avoid calling a sum a service-level measurement when the source establishes neither complete coverage nor clock agreement.
Preserve the ordering clue
Compare adjacent declarations
Inspect the public example and read each timestamp difference with its sign, offset, and source position intact.
Use a large gap to choose a log question, not a culprit
When one adjacent difference is much larger than its neighbors, identify the two declarations that bound it and their claimed receiving systems. Ask the appropriate operator which acceptance, processing, and onward-transfer events their logs can correlate. Queueing and retry behavior can produce delay, but the headers alone do not count unsuccessful attempts or reveal every intermediate wait. A field written on eventual acceptance is not a transcript of earlier conversations. Request evidence about that bounded period rather than accusing the hostname nearest the gap of holding the message for the entire interval. The useful outcome is a narrower investigation, not an invented queue duration.
Keep the Date field out of the transport-clock sequence. It is the message creator’s origination declaration, not a timestamp measured by the first relay. A draft prepared offline can have an earlier Date without having spent that whole span between mail servers. Scheduled sending, local submission behavior, and a mis-set sender clock are questions to consider, not causes this tool can establish. The From field is also a sender declaration, not independent clock evidence. For the separate meaning of address roles, read From, Reply-To and Return-Path. Do not turn a difference between Date and the first Received timestamp into a proven transport delay.
The other end of the chain also has a boundary. The last displayed Received declaration is not necessarily the moment a person saw a notification, a client synchronized, or an inbox rule moved the item. If someone reports that a message arrived at noon, ask what observation noon represents and which device or system supplied it. Compare that observation with logs only after establishing its meaning and timezone. A gap after the final supplied transport stamp may concern mailbox processing or client visibility, but this parser cannot diagnose it. Headers can narrow the period to ask about without proving final inbox placement, reading time, or the cause of a notification delay.
Keep repeated, forwarded and untrusted traces in context
Repeated Received fields are not automatically duplicate delivery. Two declarations with similar hostnames may describe internal processing stages; two identical-looking fields may instead reflect copying, an unusual system, or alteration. Preserve every occurrence and its position rather than deduplicating them to make the route shorter. A repeated id token is a correlation lead, not a globally unique proof that the same queue event happened twice. Conversely, different ids do not establish different messages across independent systems. Use the recipient administrator’s correlation rules and logs to decide what an identifier means in that environment. The local display has no access to those records and cannot verify its uniqueness.
Forwarding, list processing, and re-submission can add transport declarations while preserving older ones. That makes the oldest retained field different from a guaranteed original submission event. Comparing two delivered copies can also produce different final segments without either copy supplying the complete story. Preserve which mailbox and source each block came from before comparing them outside this single-input tool. Do not splice two blocks into a synthetic route or fill one copy’s unknown interval with the other’s date. For the distinct effects on authentication, use forwarded email authentication. Here the practical question is which declarations belong to this supplied copy and what each adjacent pair can support.
A coherent-looking route can still contain forged declarations. Neither a plausible hostname nor a convenient position proves who added a field. Authentication-Results fields likewise remain separate in source order, and a declared authserv-id is not automatically trusted. A parsed pass or fail is reported text, not fresh verification; ARC is different material and does not validate the route here. The analyzer performs no DNS lookup, SPF query, cryptographic verification, or DMARC alignment calculation, and it supplies no spam, deliverability, or safety score. Declared DKIM domain, selector, and algorithm values remain declarations. Read SPF, DKIM and DMARC header results for that interpretation task.
Prepare a precise inspection and a limited handoff
Obtain a complete header block using the Gmail headers guide or Outlook headers guide, then preserve its order when entering it. The source bounds are 128 KiB of UTF-8 text, 2000 lines, 500 fields, 128 Received fields, and 32 fields across Authentication-Results and ARC-Authentication-Results combined. The retained value of any single field must also stay within 8,192 UTF-16 units, including folded whitespace and line endings. Oversized or malformed input is rejected atomically, including orphan continuation lines and non-header text before the separator. The first empty line ends the header block; later body content is ignored, never rendered as mail or attachments. An unsupported clause stays unsupported rather than receiving an invented interpretation. Rejection is not evidence that a message was delayed or malicious.
The genuine analyzer uses an isolated opaque-origin standalone PHP document and a local terminable worker with a six-second deadline. It uses no analysis API, remote route service, storage, history, Copy command, PDF, export, or project feature. Clear wipes the input and all derived DOM and stops running work. Becoming hidden, pagehide, pageshow, and freeze also wipe the working session, so switching away does not park a result for later. This is not an absolute memory-erasure claim or protection against extensions, a compromised device, or a hostile parent substituting another interface. The containing website can have ordinary page traffic separate from the offline parsing job.
Frame the handoff around the two source occurrences, their complete declared timestamps, the signed difference when available, and the specific log question. Mark unknown offsets and negative differences explicitly. Avoid sending unrelated recipients, private host details, or message identifiers to people who do not need them; consult sharing email headers safely before disclosing source. The Email Header Analyzer hub explains the broader scope, and the Tools directory leads to other tasks. Finish with evidence-sized language: these declarations contain this gap. Actual queue residence, responsibility, and complete elapsed delivery time remain questions for corroborating records rather than conclusions supplied by a local parser.
Questions about Received Headers and Delays
Why does the route not follow ascending timestamp order?
It reverses the Received fields from the supplied source, showing the oldest retained declaration first by that ordering convention. It does not date-sort. A timestamp that conflicts with the apparent progression stays in place so the clock or ordering discrepancy remains visible.
Does a fifteen-minute difference prove a fifteen-minute server queue?
No. It is a signed difference between two usable declared timestamps, not a measured queue interval. The gap can guide a request for acceptance, retry, or onward-transfer logs, but this tool cannot attribute the whole interval to one server or cause.
Can I treat a missing offset or -0000 as my local timezone?
No. The analyzer keeps missing or unsupported time information unknown and treats -0000 as unknown timezone information, not an ordinary explicit UTC offset. It does not infer a zone from your browser or a hostname. Leave the adjacent difference unknown rather than inventing one.
Continue with another email-header guide
Interpret a supplied header in the Email Header Analyzer or choose a related task below. Explore more practical utilities in all free tools.
Ask for the missing evidence
Locate the interval to investigate
Use raw headers to identify a bounded timing question, then seek corroborating records before assigning delay.