What is connection concurrency?

Still have a question, spotted an error, or have a better explanation or a source we should cite?

Imagine your mail server trying to deliver 50,000 emails at once. It doesn't send them one at a time. Instead, it opens multiple simultaneous connections to the receiving server and sends batches in parallel. That's connection concurrency in a nutshell.

Connection concurrency is the number of simultaneous SMTP connections your sending server (MTA) keeps open to a receiving mail server at the same time. Think of it like checkout lanes at a grocery store. More open lanes means faster throughput, but open too many at once and the store manager asks you to shut some down.

The problem is that inbox providers like Gmail and Outlook set their own limits on how many simultaneous connections they'll accept from any one sender. Push past those limits and they'll reject the extra connections outright. Do it repeatedly and it starts to look aggressive, which can hurt your sender reputation.

A few things that affect how many connections you can safely open:

  • Your sending volume. Higher volume justifies more connections, but only if the receiving server agrees.
  • Your sender reputation. Established senders with clean histories often get more breathing room than new or recovering ones.
  • The destination provider. Gmail, Outlook, and Yahoo all have different tolerances, and those tolerances can change without notice.
  • Your IP warmup stage. New IPs should start with lower concurrency and increase gradually as trust builds.

If you're using an ESP like Postmark or Amazon SES, concurrency is managed for you. Their infrastructure already knows the tolerances for major providers and adjusts automatically. You only need to think about this actively if you're running your own MTA software directly.

For self-managed senders, the general rule is to start conservative, watch your delivery logs for connection rejections or 421 errors (which usually mean "slow down"), and adjust from there. There's no magic number that works everywhere. It's genuinely destination-by-destination.

Not sure how your current setup is performing? Our free Email Header Analyzer can help surface delivery signals, or drop into the SOS hotline if something's actively breaking.

Contributors

Who worked on this answer

Every name links to their profile. Every company links to their site. Real people, real accountability.

Ask an AI · tailored to your setup

Get concurrency guidance for my setup

I run my own MTA and I'm trying to figure out the right connection concurrency settings for my sending setup. Based on my situation, give me ranked recommendations for how to approach this: 1. My current sending volume per day 2. My top destination providers (Gmail, Outlook, Yahoo, etc.) 3. Whether my IPs are warmed or new 4. Whether I'm seeing 421 errors or connection rejections in my logs For each factor, tell me whether it suggests I should increase, decrease, or hold my current concurrency settings, and explain why.

Edit the yellow boxes, then send to the AI of your choice.