Skip to main content
.com domains from $2.99 — free WHOIS privacy on every name

Engineering Glossary · Mail Transport

SMTP: Sending Mail Inside a Request Is a Latency Bug

Checkout that takes four seconds to confirm is usually not a slow database; it is an order email being handed to an SMTP server while the customer waits.

The short answer

SMTP is the protocol that carries mail outwards — from your application to your mail server, and from that server onwards to others — and every one of those conversations takes several network round trips.

That matters because of where the sending happens. If your order confirmation is dispatched inline, inside the request that processes the payment, the customer waits for a TCP connection, a TLS handshake, an authentication exchange and a message transfer before the page renders. Move the send onto a queue drained by cron and the same page returns in a fraction of the time, with nothing about deliverability changed.

By the HostingFast team · Reviewed 12 August 2026

0

Jargon left undefined

100+

Entries, all cross-linked

Real

Working examples

Free

To read, always

IMAP and POP handle reading; SMTP handles dispatch. Legitimate mail joins the system through authenticated submission — your credentials over an encrypted port — which is why 'mail comes in but nothing goes out' points at an SMTP setting far more often than at a broken mailbox.

It is also where deliverability is decided. The sending server's reputation, whether SPF agrees, whether DKIM signs, and how your volume behaves over weeks all influence whether receivers trust what SMTP hands them.

The three settings that account for most failures

Authentication must be switched on for the outgoing server — the checkbox nobody ticks. The port is 465 with implicit TLS or 587 with STARTTLS. The username is the full email address, not the part before the @. Those three settle very nearly every case of mail arriving but never leaving, and they are worth confirming before anyone opens a ticket.

In WordPress, the equivalent failure is PHP's mail() function handing messages to a local sendmail path with no authentication and no signature attached. Routing the site through a genuine authenticated mailbox instead is what turns receipts that used to vanish into receipts that pass SPF and DKIM at the far end.

Why sending mail slows a page down

One SMTP conversation is a connection, a TLS handshake, an EHLO exchange, an authentication step and a data transfer. On a good day that is a couple of hundred milliseconds; on a day the mail server is busy it is considerably more. Do that inside the request that completes a WooCommerce order and the customer is watching a spinner for the duration.

Measure it rather than assume it. Time the checkout endpoint with curl -w '%{time_total}', then disable the transactional mail plugin and time it again. The gap is the cost of sending inline, and it is normally larger than anyone expects. The remedy is a queue: write the message, return the page, send on the next cron tick.

Testing the path end to end

swaks is the right tool for this and takes one line: swaks --to you@example.com --server mail.yourdomain.com:587 -tls -au user@yourdomain.com -ap. It prints the entire conversation, so a failure is attributable to a specific step rather than to 'email is broken'. Without swaks, openssl s_client -connect mail.yourdomain.com:587 -starttls smtp gets you far enough to see the banner and the advertised AUTH mechanisms.

Then verify the other end. Send to a mailbox you control, open the message source, and read the Authentication-Results header. It states plainly whether SPF passed, whether DKIM verified and against which domain. That header is evidence; a plugin's green tick is not.

Where this lands on HostingFast

Mailboxes on your own domain come with every plan, with webmail, IMAP, POP and SMTP available from the start. Hosting mail comes with sending limits and a reputation shared between tenants, so campaigns belong with an email service provider while correspondence and site notifications belong here.

Order an annual plan and the first year of your domain registration costs you nothing. Follow-on entries: SPF, DKIM, DMARC and MX Record.

Working through an inbox that carries its own domain name

A hundred terms, each with the mechanism

Every entry here started life as a support ticket. Rather than explain the same term for the two hundredth time, we wrote it down properly — with the command that measures it and the number that counts as good.

The rate you sign up at is the rate you renew at, so year two costs exactly what year one did — no ambush waiting in the invoice.

  • 100+ entries, each with the mechanism spelled out
  • Commands you can paste, numbers you can compare against
  • Neighbouring terms wired together at the foot of every page
  • Written by the engineers who answer the tickets

Why HostingFast

Standard on every plan

The whole conversation, printed

Where a protocol exchange can be dumped to a terminal, the entry gives you the command that dumps it.

Headers as evidence

Authentication-Results and Received beat any plugin's green tick, and the entries say which header to read.

Numbers rather than adjectives

Round trips, ports and timeouts are named, so 'it feels slow' becomes a figure you can compare next week.

Latency traced to its source

Where a feature sits inside the request path, the entry says so — and says how to get it out of there.

Wired to its neighbours

SMTP runs into SPF, DKIM and DMARC, because dispatch and deliverability are one problem wearing three names.

This term, landed

SMTP defined, its ports fixed, its cost measured and its test written out — enough to debug a send tonight.

Quick Start

From order to online

  1. 1

    Fix the submission settings first

    Authentication on, port 465 with TLS or 587 with STARTTLS, full address as username. Most send failures end here.

  2. 2

    Dump the conversation with swaks

    swaks --to you@example.com --server mail.yourdomain.com:587 -tls -au user@yourdomain.com -ap prints each step, so the failure has an address rather than a vibe.

  3. 3

    Get the send out of the request path

    Time the endpoint with mail on and mail off. If the gap is material, queue the message and drain it on cron rather than making the visitor wait for it.

Built In

Loaded onto every plan

  • Mailboxes on your own domain, included with the plan
  • Browser webmail plus IMAP, POP and SMTP for any mail client
  • NVMe SSD storage on every tier, not only the expensive ones
  • Zero setup charges — there is no joining fee, ever
  • Daily backups with self-service restores you run yourself from the panel
  • Free SSL on every plan, renewed automatically before it can lapse
  • Per-site PHP version switching from the control panel
  • WordPress and 400+ other applications installed in one click
  • Human support on duty every hour of every day
  • WebP image conversion built in, at no extra cost

Frequently Asked

What people ask us most often

Which submission port and encryption should an application use?

Port 465 with implicit TLS, or port 587 with STARTTLS. Both are correct; pick whichever your library handles more cleanly. Authentication must be enabled on the outgoing server and the username is the full email address. Port 25 is for server-to-server relay, not for your application, and plenty of networks block it outbound anyway.

Why does checkout get slower when order emails are switched on?

Because the message is being handed to the mail server inside the request. That is a TCP connection, a TLS handshake, an auth exchange and a transfer, all with the customer watching. Time the endpoint with the mail plugin on and off to size it, then queue the send and drain the queue on cron. Deliverability is unaffected; only the waiting moves.

What sending volume is hosting SMTP built for?

Correspondence and site notifications. Mail on hosting carries sending limits and a reputation shared between tenants, and bulk traffic degrades that reputation for everyone on the machine. Newsletters and campaigns belong with an email service provider; there is no configuration that makes shared hosting the right dispatcher for them.

Is there a credit if a month misses the uptime target?

Yes. The target is 99.9%, and if a month drops below it because of a fault on our side, our terms entitle you to a pro-rated credit — ask and we apply it. We deliberately call it a target rather than a contractual SLA. Hardware and network faults surface through platform monitoring, usually before the first customer notices anything.

Keep reading

  • IMAP

    The reading half of the same stack, and where folder shape decides how fast a client feels.

  • DKIM (DomainKeys Identified Mail)

    The signature that survives a forwarding hop, and the header you read to prove it verified.

  • How to Test Your Website Speed

    Measuring page speed properly — the same habit, applied to the request rather than the message.

  • VPS Hosting

    KVM virtual servers — root access, DDoS filtering and one flat monthly price.

  • WooCommerce Hosting

    Store-ready hosting with Redis caching and headroom reserved for checkout traffic.

Changing hosts? Run through our checklist first.

A straightforward sequence for a switch your visitors never feel: which files move first, how to shift email across without losing a single message, the right moment to repoint DNS, and the two mistakes behind almost all the downtime we get asked to rescue.

You'll get the checklist email, then occasional pointers on keeping a site running fast. Unsubscribe the moment you want out — the privacy policy covers the rest.

Take the send out of the critical path.

NVMe on every tier, cron you control, mailboxes included — and renewals billed at the rate you originally signed up at.

View VPS Hosting plans