Do we have rate-limits for SMTP?

Like the API, SMTP rate-limiting allows us to ensure resources are available for all accounts, and that a single account cannot significantly impact access for other customers. However, our limits for SMTP are somewhat arbitrary, and informal.

SMTP relaying is also inherently a little less prone to failure from client code - many customers have an smtp server on their network that will queue mail, and relay it to Postmark asynchronously. This means that some customer code is slightly more tolerant to SMTP downtime than API downtime.

We recommend that customers limit the total number of connections to our SMTP endpoint to 10 concurrent connections, per IP. For SMTP, connections can be reused, and more than one message can be submitted within the same connection.

We do not have any systems in place to enforce these limits, we rely on customers to self-police and be good actors on our systems.

However, this is currently being discussed here by the Postmark team.

At some point in the future, we may enforce connection limits, but this is a much more complex problem, as there may be no way to do this on a per-server, or per-account basis due to the load balancer we use, and the nature of the SMTP protocol.