Slow filtering on WASieveServer on the Broken encoded message

Indreias

In continuation of this thread.
Axigen TNEF Decoder (10.2.3.2 )

I found out that the delay is coming from “WASieveServer”.

For this example in which I only tested on a “12kb WASieveServer file”.
And it took 1.5mins to finish the processing for a “Broken Encoded message with 5.9mb size”.
2022-05-27 13:03:55 +0400 08 axigen PROCESSING:003CE5EF: Start filter WASieveServer of type script filter from server
2022-05-27 13:05:34 +0400 08 axigen PROCESSING:003CE5EF: Keep requested for XXX@XXX.com
2022-05-27 13:05:34 +0400 08 axigen PROCESSING:003CE5EF: Finished filtering mail object 3CE5EF with filter: WASieveServer of type script filter from server

This became a problem to one of our deployment whom received the broken encoded message, which has 250kb WASieveServer file.

With this basis its accurate what I said before that it needs 30-45mins to release if from queue. ((250 kb / 12kb) * 1.5mins))

And for a Good message with a size of 12mb it took seconds only to process which is good due to the size of it.

Hope you can give me Idea how to solve it.
This issue is not frequent but it happen.

Regards,
Jay

Hello Jay,

Thank you for your message.

In order to better understand your current situation, please attach to your response the server SIEVE rules file available by default on Linux systems at:

/var/opt/axigen/filters/wasieve-server.sieve

Thank you.
Bogdan Maxim

@MaXiM

I will message you privately the wasieve-server.sieve.

Regards,
Jay

@MaXiM

I found the cause and it was the wasieve-server.sieve. Due to big filters we have on the server and also some are checking into body of the message and this is the reason why it takes too long to filter.

It tries to read all the broken encoded message on the body.

I will try to eliminate some of this filter to atleast improve the situation.

Regards,
Jay

Hello Jay,

Thank you for your response.

Please know that we have noticed that you have 4 different rules that have as condition to search the message body for a specific text.

We would like to mention that when using sieve rules to filter the body of the messages it is strongly recommended to perform the filtering only if the message size is small because for (very) large messages this type of filtering is expensive in terms of resources and can cause slowdowns in the processing of the entire queue.

To do so, you may add an additional condition to each rule so that the rule applies only if the message size is lower than a certain value (10 KB) in our example. We attach a screenshot. It is important to select “ALL of the conditions below” at the top of the Conditions section.

@MaXiM

Thanks for the assistance and we both came to similar conclusion.
End of incident

Thanks and Regards,
Jay