SMTP 421 error: temporary failure or real problem?

Quick answer for smtp 421 error

IETF RFC 5321 — Simple Mail Transfer Protocol classifies 4xx as transient. The sender or the forwarder should retry. 5xx is permanent. Mixing those classes is how people restore Workspace MX because Gmail asked for ten minutes. Email forwarding rejection diagnosis names the hop first: prove who rejected. This page is what to do when that hop said 421.

Temporary: one 421, later 250, exclusive MX unchanged. Real problem: 421 for hours, same enhanced status, or 421 used as a polite permanent refuse. Also real: you never had 421—you had empty history and called it 421. Read the line.

Related temporary codes already explained on the site: 450 vs 451 vs 452. 421 is the “closing or not available” cousin. Do not merge a 535 AUTH ticket into 421. AUTH is outbound client login.

MailerZ Free is one domain, three aliases, 14-day store, unknown held, no send-as. Paid store is 90 days. Solo is $40 per year. Starter is $8 or $80. Business is $19 or $190. Agency is $39 or $390. Unlimited is $99/month or $990/year. Confirm pricing. The store lets you see whether 421 became 250. It is not an inbox SLA.

Probe from a mailbox you do not own. Self-send lies. Exclusive MX first. Leftovers are not 421.

The user problem and the decision criteria

Founders see 421 in a bounce and assume the product is down. They add the old MX. Now some mail is 421 at Gmail and some mail is on Google leftover. The smtp 421 error they had was hop two asking for time. The outage they created is hop one split.

Agencies write “421 at HH:MM, dest=gmail, no DNS change” in the channel. That sentence prevents a helpful contractor from republishing ASPMX.

421 can be greylisting, overload, or a policy that looks temporary. You cannot know from the number alone. The enhanced text and the retry outcome tell you. One retry success is temporary. A day of identical 421 is real.

Technical mail flow when 421 appears

Hop one: sender to MailerZ MX. Named alias, store, 250, SRS on envelope, Header From intact. Email forwarding. If hop one never happened, you will not see 421 from Gmail in MailerZ history.

Hop two: MailerZ submits to Gmail or Outlook. 421 means that server closed or deferred. A well-behaved client retries. History should show the later attempt. If the later attempt is 550, the problem graduated to permanent. Different ticket.

IETF RFC 5321 — Simple Mail Transfer Protocol does not give you a wall-clock SLA for how long to retry. Operators retry with backoff. Do not interpret “not in the inbox in three minutes” as 421 failure. Watch the hop.

Free hold is not 421. Unknown local-parts are held. Do not call hold a temporary destination error. Create the alias or read hold.

Recovery is 14 days on Free and 90 on paid. That window is how you prove 421 became 250 yesterday. It is not an archive. MailerZ is not IMAP.

Outbound send-as 421 from MailerZ SMTP is rare in this inbound article. If a WordPress plugin shows 421 while submitting, that is a different host conversation. Copy that pair from the dashboard. Do not change inbound MX.

When 421 is still temporary

1. The line is actually 421

   Not 550, not empty history, not a Gmail spam folder. Copy it.

2. Hop two returned it

   MailerZ already accepted. If not, you are in leftover MX or missing alias.

3. Exclusive MX still exclusive

   Two resolvers. No Google leftover. Leftover MX.

4. You wait one retry cycle

   Minutes to a short hour, not a weekend of toggles. Re-read history.

5. Later 250 closes it

   Temporary. Write the pair of timestamps. Do not “fix” SPF.

When it is real: same 421 text across many attempts, destination account disabled, or 421 used as a tarpit. Then you escalate to the destination admin with the copied lines. You still do not add a second MX.

450, 451, and 452 remain sibling 4xx codes. If history shows those instead, use that article. Do not rename them 421 in Slack.

Step-by-step: handle a 421 without making it worse

1. Reproduce with a unique subject

   Third mailbox. Named alias. Exclusive MX. Self-send off.

2. Copy hop two

   421 plus enhanced text plus UTC. If the row is empty, leave this page.

3. Freeze DNS

   No leftover add. No preference games. MX priority is not load balancing.

4. Watch history

   Later 250: done. Later 550: permanent now. Still 421 after hours: dest problem.

5. Tell the customer the class

   “Temporary deferral at Gmail, we are not changing MX.” That sentence prevents dual MX courtesy.

6. Only then escalate dest admin

   Attach lines. Do not attach a theory about SPF.

If send-as also failed the same hour, split tickets. Inbound 421 and outbound 535 are not one outage. Free still has no send-as.

Weekend 421 is still 421. An on-call who adds Google MX at 11 p.m. creates Monday leftover. Schedule the human who may not edit DNS.

Failure modes people call 421

Empty hop. Proof: no row. Fix: leftovers or sender. Not 421.

Hold. Proof: unknown local-part. Fix: create alias. Not 421.

Spam folder after 250. Proof: delivered, filtered. Fix: not a 421. No inbox percentage.

Toggle storm. Proof: MX edited during wait. Fix: exclusive restore. Sit.

Misread 550 as 421. Proof: the line says 5xx. Fix: destination permanent checklist.

Outbound plugin 421. Proof: WordPress log, not inbound history. Fix: SMTP pair on a paid plan.

Sticky office resolver plus leftover. Proof: public views exclusive, one LAN not. Fix: that LAN, not dual MX.

Store expired. Proof: you look for last month’s 421 on Free. Fix: 90-day store next time. The wait still should not have included a second MX.

From rewrite at another vendor. Proof: Header From moved. Fix: leave that hop. MailerZ does not rewrite From. Unrelated to 421.

Null MX leftover. Proof: refuse-all published. Fix: delete null MX. That is hop one refusal, not dest 421.

MailerZ workflow and the product boundary

MailerZ retries hop two when the destination returns 4xx, within the product’s retry behavior. It does not promise a minute. It is not Workspace, not IMAP, not an open relay, not SOC 2. Security. Docs.

Seats operate the dashboard. They do not shorten Gmail 421. Agency does not buy a faster dest.

Header From intact. Envelope SRS only. A 421 after an honest hop is still hop two. Do not rewrite From to dodge it.

Compare: Cloudflare Email Routing, ImprovMX. Ask whether they store dest SMTP. This page needs that store.

Unhosted or unauthorized send-as is 550, not 421. Keep the numbers honest.

Cost of waiting versus the cost of a courtesy leftover

Waiting is free. Dual MX is not. Solo at $40 per year keeps 90 days of hop lines so you can show 421 then 250. Starter, Business, Agency scale domains. Annual paid except Solo includes two months free versus monthly. Confirm pricing.

Do not buy Google Workspace — product overview because of one 421. Buy a suite when you need Calendar, Drive, and hosted stores. Not because Gmail asked for a retry.

Staff time: one person owns the wait. Mute “is it done” for the retry window. Put UTC in the channel.

Tools: tools, MX lookup, leftover MX. They prove hop one. 421 lives in history.

Legal still does not get a seven-year archive from a 421 retry. Say so.

If 421 is daily for one destination address, fix that mailbox. Paying Agency will not change Gmail’s reply to a disabled user.

A retry window is cheaper than a leftover host you will hunt for a week. Write the customer the class of error. Temporary means they may see a delay. It does not mean they should send to a second published address. Printed second addresses are how you get two stores and no owner.

Contractors who “just lower TTL and add Google back” during 421 are the usual Monday ticket. Put this page in the runbook. If they cannot wait twenty minutes, they should not have the DNS password. The smtp 421 error is a reply. It is not a vote to dual-publish.

Compare 421 to 450 mailbox busy, 451 local error, and 452 storage. All are 4xx. All want a retry. None want a second MX. If history flips between 421 and 450, you still wait. If it flips to 550, you stop waiting and read the permanent text.

Greylisting is a policy that first-time senders get 4xx and retries get 250. MailerZ as hop two may look “new” to a destination that greylists. That is still temporary. The second attempt is the test. Do not interpret the first 421 as a blocklist you must fix with SPF theater.

Rate limits at the destination can look like 421 or 4xx with “try later.” MailerZ inbound forwarding is not a campaign sender. If you blasted a purchased list through paid SMTP, you left the product boundary. Hourly send-as caps exist on paid cards. They are not an excuse to spray. This inbound 421 page assumes a person sent one message to a named alias.

IPv6 versus IPv4 retries can differ if a destination is sick on one family. You will see that in the hop text if the product shows the connecting attempt. Still do not add leftover MX. Fix or wait the dest. If you cannot read the family, you still have the class: 4xx retry.

Related: deferral versus permanent when that page exists, troubleshooting, delivery recovery. Use recovery for hops MailerZ took. 421 then 250 is a success story in that store. 421 then empty because you deleted leftovers incorrectly is not.

Write the on-call rule: no MX edits during a documented hop-two 421 unless leftover MX is already proven on two resolvers. Those two facts are almost never true at the same time. If leftovers exist, you never had a clean 421 ticket. Delete leftovers first. Then see if 421 remains.

Solo, Starter, Business, and Agency do not include a 421 accelerator. The dashboard shows the line. You supply the wait. Free can show the line for 14 days. That is enough for a retry cycle. Upgrade for store length and send-as, not for patience.

If legal asks why a message was late, the answer is destination 421 at UTC timestamps, then 250. That is a complete story. “We added Google MX to be safe” is how you lose the story and the next invoice.

Field notes when 421 keeps showing up

Put 421 in the ticket title with the dest hostname and UTC. “Email slow” is not a title. The next on-call should see hop two without a transcript.

If only one destination address 421s and others 250, you have a mailbox problem. Do not change MX for one disabled user. If every destination 421s, you still wait once, then escalate dest-wide, still without leftover MX.

Providers sometimes 421 during their own incidents. Your hop line plus their status page is enough. Adding Google MX during their incident guarantees you will forget to delete it.

WordPress “could not send” banners are not inbound 421. Read the plugin log. If it is AUTH, leave this page. If it is 421 to MailerZ SMTP on a paid plan, copy the dashboard pair and retry. Do not touch inbound MX.

Mobile users who never saw the delayed message after a later 250 should search all mail. The retry succeeded. Their swipe archive is not a 421.

Agencies: one 421 on one client is not a fleet event. Do not “fix” ten zones because one dest greylisted. Parallel MX edits during a dest blip are how you earn ten leftover tickets.

Write the customer: “Gmail asked us to retry. We are not changing DNS. You should see the message after the next hop.” If they demand a second address, give them a named alias that already exists, not a second MX.

If hop history shows 421 then nothing because the store window expired, you lost the second timestamp. Paid 90-day store exists so last week’s retry is still visible. Free 14 days is enough for a same-day 421. Upgrade for length, not for magic. Solo at $40 per year is the usual card when you want that window and send-as later.

Do not run a load test to “prove 421.” You will earn real 4xx and look like a campaign. One unique personal message is the test. MailerZ is not a bulk ESP. Hourly send-as caps on paid plans are for operational mail, not for hammering Gmail until 421 disappears.

Document dest 421 in the same runbook as leftover MX so people stop mixing them. One page for hop ownership, one page for 421 class, one page for MX-to-inbox path. Three titles. Three owners. That is how a small team stays sane.

How to read 421 enhanced status text

The three-digit class is the first filter. The rest of the line is the ticket. Copy the enhanced status if it is present: strings such as 4.3.2, 4.4.2, or 4.7.0 are not decoration. They tell you whether the dest claimed system not accepting, a routing timeout, or a policy deferral. One hop that says “service not available, closing transmission channel” is a close. One hop that says “too many connections” is capacity. Both are still 4xx. Neither is leftover MX. Treat the text as evidence you will paste, not as a slogan you rewrite in Slack.

Connection-time 421 is different from 421 after MAIL FROM or RCPT TO. If the dest closed before it accepted the envelope, MailerZ history should still show an attempt and a deferral if hop one already succeeded. If you never see hop one, you do not have dest 421. You have a sender that never offered the message to MailerZ. That is leftover MX, a missing alias, or a queue on a host you do not operate. Do not diagnose a close banner you never stored.

Some destinations emit 421 with almost no English. Keep the numeric class, the enhanced code, the hostname in the banner, and the UTC from history. Guessing “Gmail is down” from a three-digit number wastes the dest admin’s first reply. Guessing “we need SPF” wastes yours. SPF does not emit 421. DKIM rotation during a single deferral does not clear 421. Those records matter for send-as and for later reputation work. They are the wrong lever while hop two is asking for a retry.

A timeout with no SMTP reply is not 421. Operators blur “could not connect” and “421 try later” because both feel temporary. They are not the same artifact. A connect failure is a path or dest listener problem. A 421 is a spoken reply. If history shows neither a 421 nor a 250, you may have a timeout class. Write that honestly. Adding Google MX still does not repair a dest that is not answering. It only splits hop one.

Multiple destinations in one hour need separate lines. Gmail 421 on one alias and Outlook 250 on another is not a MailerZ outage. It is two hop-two conversations. Agency seats do not merge those replies. Solo at $40 per year does not either. The store lets you keep both timestamps. Free keeps them for 14 days. Paid keeps them for 90. That is the product fact. It is not an inbox percentage and not a dest SLA.

Greylist, dest incident, and polite refuse

Greylisting is a policy: first-time or unfamiliar submissions get 4xx, a later retry gets 250. MailerZ as hop two can look new to a dest that greylists forwarders. The smtp 421 error on attempt one is then expected. Attempt two is the test. If the second hop is 250 and the message appears, you waited correctly. Do not open a blocklist ticket. Do not republish ASPMX. Do not rewrite Header From. MailerZ rewrites envelope with SRS only. From stays intact. A greylist is not a reason to change that boundary.

Dest incidents are louder. Status pages, many aliases 421 at once, and the same enhanced text across customers point at the dest, not at your zone. Your job is still exclusive MX and copied lines. Adding a leftover “until they recover” guarantees you will forget the leftover. When their incident ends you will have two owners and empty history on the copies that went to Google. That is how a temporary dest event becomes a week of leftover MX. Put the dest status URL in the ticket. Do not put a second MX host in the zone.

Polite refuse is the ugly cousin. Some operators use 421 as a soft 550: same text for days, no later 250, mailbox gone or policy permanent. The class is still 4xx on the wire, so clients keep retrying. You escalate with the repeated line, not with a theory. Hours of identical 421 is a real dest problem. A day of it is still not leftover MX. The dest admin or the mailbox owner has to change something on their side. MailerZ will not turn that 421 into a 250 by selling Agency.

Tarpits and connection-rate 4xx look like 421 or sibling codes when someone sprayed. This inbound page assumes a person sent one message to a named alias. If you used paid SMTP to blast a list, you left the product. Hourly send-as caps on paid cards exist for operational mail. They are not a campaign license. Free has no send-as. Unauthorized send is 550 / 550 5.7.1, not 421. Keep those numbers in different tickets.

IPv4 versus IPv6 can split a dest that is sick on one family. If history shows the connecting attempt and the family, keep that in the paste. You still wait or escalate the dest. You still do not add leftover MX. If you cannot see the family, you still have the class: 4xx retry. Do not invent a dual-stack MX “backup” that is actually Google’s leftover set.

Related reading already on the site stays specific: 450 versus 451 versus 452 for sibling 4xx, who rejected for hop ownership, leftover MX for hop one. This page owns the 421 class only. If the line is 550, leave. If the row is empty, leave. If the line is 421, stay and wait once.

The paste packet that stops leftover MX

On-call fails when the channel says “email slow.” Write a packet a contractor can refuse to “fix” with ASPMX. Include: domain, exclusive MX confirmed on two public resolvers, alias that was offered, hop-one 250 timestamp, hop-two 421 full line, enhanced status, dest hostname, UTC, and whether MX was edited after the first 421. If MX was edited, say so first. That ticket is no longer a clean 421. Restore exclusive hosts, then see if 421 remains.

Customer language is short. “Gmail asked us to retry. We are not changing DNS. You may see a delay. We will re-read history after the next hop.” If they demand a second printed address, give a named alias that already exists. Do not give them a second MX operator. Printed second addresses become two stores and no owner. Hold is still not 421. If they wrote a name you never created, create it or read HOLD. Do not wait on dest 421 for a hop that never left the hold queue.

Agencies isolate one client. One dest greylist is not a fleet MX edit. Parallel zone changes during a dest blip are how you earn ten leftover tickets on Monday. Put this page and the leftover-MX tool in the handoff. The person with the DNS password should not be the person who is bored during a twenty-minute wait. If they cannot sit, they should not have the panel.

Legal asks why a message was late. Answer with destination 421 at UTC, then 250, plus exclusive MX unchanged. That is a complete story. “We added Google to be safe” deletes the story. The 14-day Free store is enough for a same-day pair of timestamps. Upgrade to paid 90-day store when last week’s retry must still be visible. Upgrade for send-as and limits when those are the job. Do not upgrade to buy patience. Confirm pricing the day you quote. MailerZ is not IMAP, not Workspace, not an open relay, not SOC 2.

Weekend 421 is still 421. Night leftover MX is a Monday outage. Schedule the human who may not edit DNS. Mute “is it in yet” for one retry cycle. If later history is 550, graduate the ticket to permanent dest text. If later history is 250, close. If later history is still 421 after hours with the same enhanced line, escalate dest with the packet. If history is empty, you were never on this page. Query leftovers. Create the alias. Probe from a mailbox you do not own. Self-send still lies.

Do not load-test to “prove 421.” One unique personal message is the test. MailerZ inbound forwarding is not a bulk ESP. Hammering Gmail until the number disappears is how you earn real 4xx and look like a campaign. The smtp 421 error is a reply. It is not a vote. It is not a status badge. Wait, copy, classify, then stop.

FAQ

Key takeaways

• 421 is 4xx. Retry, do not toggle MX.
• Name the hop. Empty history is not 421.
• Accept 250 plus dest 421 is hop two.
• Later 250 means it was temporary.
• Hours of identical 421 is a real dest problem.
• Leftover MX is a different ticket.
• Hold is not 421.
• Do not rotate SPF during one 421.
• Self-send can hide the hop.
• 450/451/452 are siblings, not synonyms.
• Free store 14 days. Paid 90.
• MailerZ is not IMAP, not a suite, not SOC 2.
• Courtesy leftover MX turns a wait into an outage.

Conclusion and next action

An SMTP 421 error is temporary until history proves otherwise. The real problem is usually leftover MX, a missing alias, or a destination that keeps saying 421. Dual MX is never the fix.

Next action: copy the line, confirm exclusive MX, wait one retry, re-read history. If 250, stop. If still 421, take the enhanced text to the destination admin. If the row is empty, you were never on this page.

Keep the two timestamps. The next 421 should be boring. That is the goal. If the second timestamp never arrives, you still have a dest problem or you never had hop two. Re-classify with the rejection page. Do not invent a third MX host while you think.

Review quarterly, or sooner if hop status labels change. Author: MailerZ editorial, Secuno LLC.