Inbound and outbound SMTP rate limiting is present, even when none is configured

[Morpheus0x]

I am using Stalwart to send out a newsletter to about 30.000 recipients using Listmonk. At first, I had an issue with the limit for sending emails in a single SMTP session. After some searching, I was able to solve this by adding the following configuration to config.toml:

[session.data.limits]
messages = 10000
size = 104857600
received-headers = 50

After doing that, I didn't get any errors in Listmonk anymore. However, the message queue in stalwart grew steadily. Thinking that this probably was an issue with the recipient server, I didn't investigate further. Until I noticed some emails from [email protected] to [email protected] were also in that queue, with the server response being Error for localhost: Rate limited.

This confused me, since I deliberately disabled any inbound and outbound rate limitings under the SMTP settings. By that time, I already had about 9.000 emails from MAILER-DAEMON in the [email protected] inbox.

After looking at any other rate limiting settings and reading the docs twice, I didn't find any mention of any other applicable rate limitings or default limits. In desperation, I added the following config as described in the docs for both inbound and outbound SMTP:

[[queue.limiter.outbound]]
rate = "100/1s"
enable = true

[[queue.limiter.inbound]]
rate = "100/1s"
enable = true

After restarting stalwart, my inbox instantly grew by another 10.000 emails from MAILER-DAEMON, which in my eyes proves that there are some default rate limits applied, even when none are configured. Now I don't see any emails from MAILER-DAEMON in the queue anymore, and they seem to get delivered to the newsletter@ inbox much more frequently.

Can someone explain why this default, undocumented rate limiting exists? Is this a bug or intended design?

As a sidenote, I found some UI bugs for the message queue: The maximum items in the queue that can be displayed in the UI is limited to 100, but the actual queue size seems larger. After restarting the stalwart server, the next retry timestamp gets set to the current time.

[mdecimus]

It looks like you've added entries manually to the toml file. Check the local vs database settings documentation as if you don' make those setting local they will cause conflicts with the ones in the database.
When Stalwart starts you should see a warning indicating that database settings were defined in the local toml.