I guess Zammad should be in heavy load while saving the answer after editing.
I found this ticket with the same problem, but in ticket-articles:
MrGeneration posted a Skript there for manually processing all failed articles:
affected = Ticket::Article.where("body LIKE '%This message cannot be displayed due to HTML processing issues.%'")
affected.each {|article|
parsed = Channel::EmailParser.new.parse(article.as_raw.content)
# comment out if you don't want to see outputs
p parsed[:body]
# uncomment AFTER dry run
# article.update!(body: parsed[:body])
}
puts "#{affected.count} articles are affected."
Of course the command won’t work for knowledgebase answers. It’s for articles.
What happens if you try to edit? Do you see the original content or is it showing the same message?
Do you happen to know what you or the user that created the answer did try to provide?
This message appears after Zammad tried to normalize the HTML content for 20 (!) seconds.
So this may be several issues. This happens quite rarely - never happened on knowledgebase up to now.
The user tried to provide a Admin-Manual for a Software. 1-2 Word-Pages I guess(as size-measure). So text/pictures as we used often, nothing special he said.
Crazy, 20 seconds… The Knowledge Base is indeed pretty slow often. (Browser shows messages like “You really want to close this page, changes get lost” when Updating a Answer)
Zammad runs on 3 cores of our newest server (with SAN-connection)
Just now user told me that he used ~7mb images taken from his phone… I guess it was too much for Zammad at that moment. Do you know a possibility to reprocess the html content or is it lost? Just so I can tell the user if he needs to rewrite it.
If the exact content isn’t too sensitive I’d love to provide an upload space for internal testings to see if we can improve stuff at that point. Apart from 7 whopping megabyte images 2 pages of text shouldn’t™ be any issue.
I’m afraid the user has to rewrite the stuff.
As far as I’m aware we have no “RAW” stuff going on here like we do in mail contexts.
Btw in terms of image sizes:
Especially depending the type of image we had improvements on Zammad 4.1 - possibly some of your issues may vanish with an update.
I updated to Zammad 4.1 and gave the virtual machine a fourth vcpu.
Because we are just 12 Agents the WEB_CONCURRENCY thing wouldn’t help us I guess.
The User is rewriting his documentation. I will give you an update when he is ready, so we can upload it.