What Is SMTP Throttling?
TL;DR
SMTP throttling is a limit on how much mail a server will send or accept in a given period, applied either by your own provider as a quota or by the receiving provider as a rate limit. Both report it as a temporary 4yz rejection rather than a permanent failure, so the affected mail queues and retries.
Who throttles, the sender or the receiver?
Both, for different reasons. On the sending side the cap is contractual and published: Google states a limit of 2,000 messages a day for a paid Google Workspace user, and 500 a day on a trial account, measured over a rolling 24 hours rather than reset at midnight. Exceeding it suspends sending temporarily while incoming mail keeps working.
On the receiving side the cap is reputational and unpublished. A large provider slows a sender whose volume or pattern is new to it, which is why the same daily total that flows on a warmed domain gets deferred on a cold one.
What does throttling look like in your logs?
A 4yz reply with text naming the reason: rate limited, too many messages, try again later, or a policy URL from the provider. RFC 5321 defines these as transient failures, so the correct behaviour is to keep the message queued and retry rather than to report it upstream as undeliverable.
Read the distribution rather than individual lines. Deferrals spread across many receivers point at your own provider quota, while deferrals concentrated at one large provider point at that provider throttling you specifically, and the two need opposite responses.
Frequently asked questions
It raises the sending quota and does nothing for the receiving limit, because reputation attaches to the domain as well as the mailbox. Spreading the same volume for the same domain across ten new mailboxes changes the envelope and not the signal the receiver is reacting to.