A newsletter is the highest-risk mail a hosting customer sends. It goes to many recipients at once, it is the mail most likely to be reported as spam, and on a MailBaby relay it is the mail most likely to hit the 6,000 messages per hour per sender cap. None of that makes it unsendable; it makes it something to set up deliberately rather than leave to a plugin’s defaults. These are the rules we give customers who want to send lists through a relayed hosting server, and the settings that enforce them.
Table of Contents
Short answer: MailBaby discards anything above 6,000 messages per hour from one sender address, so throttle at the source: set a per-domain hourly limit in cPanel with deferral enabled, cap the list tool’s own rate, and keep the combined figure well under the limit. Send from a dedicated address on its own subdomain with SPF and DKIM, enforce confirmed opt-in and bounce removal, add one-click List-Unsubscribe headers, warm a new address up over a week, and watch the portal log during every large send so a run of rSPAM rejections stops the job early.
Respect the hourly cap at the source
MailBaby discards messages beyond 6,000 per hour from a single sender address. Discarded means gone: not queued, not deferred, and only visible in the portal log. So the sending system must never present more than that in an hour, and in practice should stay well below it to leave room for the customer’s ordinary mail from the same address.
On cPanel, set a per-domain hourly limit and let Exim defer the rest into the next hour:
whmapi1 set_tweaksetting key=maxemailsperhour value=500
whmapi1 modifyacct user=NEWSLETTERUSER MAX_EMAIL_PER_HOUR=3000
whmapi1 set_tweaksetting key=maxemailsperhour_defer value=1
The deferral setting is the important part; without it Exim rejects over-limit messages back to the script. The mechanics are in the outbound limits guide. On DirectAdmin the per-user limit is max_emails in the user’s configuration and the server default is in /etc/virtual/limit. On Postfix, smtp_destination_rate_delay = 1s paces deliveries to the relay.
List software has its own throttles. Mailman’s SMTP_MAX_RCPTS and delivery throttling, Sendy’s per-second rate in its cron settings, and MailPoet’s “emails per hour” option in its sending method should all be set so the combined rate stays under the server cap. A customer running two tools from the same address needs the sum to fit.
Use a dedicated sending address and domain
Send newsletters from an address that sends nothing else, ideally on a subdomain such as news.example.com with its own SPF and DKIM. The per-address cap then belongs entirely to the newsletter, a complaint problem stays on the subdomain’s reputation, and the customer’s transactional and personal mail keeps its own standing. Publish include:spf-c.mailbaby.net for the subdomain and a DKIM key for it; the DNS steps are in the SPF guide.
List hygiene is not optional
MailBaby scores by content and by the sending address’s history. A list containing addresses that bounce, spam traps, or people who never asked to receive the mail produces bounces and complaints that lower the score of every subsequent message. Enforce:
- Confirmed opt-in for every subscriber. A signup form without confirmation collects typos and malicious submissions that become spam traps.
- Remove hard bounces after one attempt and soft bounces after three. Every list tool can do this; make sure bounce processing is actually configured, which means the bounce address must be a real mailbox the tool reads.
- Never import a list the customer bought, scraped or “found”. That is cold email and is prohibited under the acceptable use policy; it will get the address blocked and can get your account suspended.
- Re-confirm or drop subscribers who have not opened anything in a year.
Headers and content that pass
Every list message needs a List-Unsubscribe header with both a mailto: and an https: target, plus List-Unsubscribe-Post: List-Unsubscribe=One-Click. Large mailbox providers require this for bulk senders and it dramatically cuts complaints, because recipients unsubscribe instead of reporting. Confirm the sending tool adds them; most do, some only when a specific option is enabled.
Content that scores badly at the relay is the same content that scores badly everywhere: subject lines in capitals or with excessive punctuation, a body that is one large image, links through URL shorteners, link text that differs from the target domain, and a From: name that does not match the domain. A plain-text alternative part should always be present. Keep the message under a hundred kilobytes.
Warm up a new address
A new sender address on a MailBaby account, and a new MailBaby account itself, is scored more strictly until it has history. Start a new newsletter at a few hundred messages an hour and double daily over a week rather than sending 20,000 on day one. The relay and the recipient providers both read a sudden spike from an unknown address as a compromise.
Watch the send
During a large send, keep the portal log or an API query open and filter on the sending address. What you want to see is a steady stream of delivered entries with low scores. What tells you to stop is a run of rSPAM rejections, a rising bounce count, or discards for exceeding the rate. Pause the tool, fix the cause, and resume; continuing a send that is being rejected burns the address’s reputation for no benefit. The rSPAM guide covers the usual causes.
Verify
Before the first real send, send the newsletter to a seed list of addresses at the major mailbox providers and check where it lands and what Authentication-Results: says. In the Exim log, confirm the messages are leaving at the expected rate:
grep 'news@example.com' /var/log/exim_mainlog | grep -c '=>'
Run that a few minutes apart to compute the hourly rate. The common pitfall is a list tool configured to send everything in one batch through a local queue with no throttle, so the server-side hourly limit defers thousands of messages, the tool reports success, and the messages trickle out over the next several hours while the customer complains that the newsletter was late.
MailBaby newsletter sending limit at a glance

Official documentation: RFC 5321 (SMTP), Linux man pages.
Related guides: MailBaby with Plesk and Webuzo: outbound relay configuration · What is MailBaby? Outbound SMTP relay pricing, limits and how it works · How MailBaby handles feedback loops and abuse reports, and what to do when you get one.
Frequently asked questions
What happens if a customer sends more than 6,000 emails per hour through MailBaby?
The excess is discarded, not queued or deferred, and the only record is in the portal log. The customer sees a partial send with no local error, which is why the throttle has to live on your server or in the list tool rather than relying on the relay.
Can I send a purchased or scraped mailing list through MailBaby?
No. Cold email to recipients who did not opt in is prohibited under the acceptable use policy, produces the complaints and spam-trap hits that block the sending address, and can get the whole account suspended. Only confirmed opt-in lists with working bounce processing should go through the relay.
How long should I warm up a new newsletter address on MailBaby?
Start at a few hundred messages an hour and roughly double each day over about a week. Both the relay and the large mailbox providers treat a sudden spike from an address with no history as a likely compromise.