Nagios Support Forum

Hi all,
I'm hoping someone can point me in the right direction. I've been running Nagios XI for a few months without major issues, but recently, I've noticed that email notifications are arriving much later than expected. The alerts are generated correctly in the interface, and the notification history shows they were sent immediately, but the emails don't actually reach my inbox until 20 to 30 minutes later.
What's confusing is that this only affects notifications from Nagios XI. Other systems using the same SMTP relay deliver messages within seconds. I've checked the mail queue on the XI server and don't see anything building up there. CPU and memory usage are normal FNF, and I haven't made any major configuration changes recently except applying a routine system update a couple of weeks ago.
I also tested sending a message directly from the command line on the Nagios server, and that arrived almost instantly. Because of that, I'm struggling to determine whether the delay is coming from Nagios itself, the notification subsystem, or something related to how XI is handing messages off to the mail server.
Has anyone run into delayed notifications like this before? I'm interested in knowing which logs or services I should focus on first.

Hello @Mikasa23!

With what I have observed I think what may be happening is the event queue is being loaded with events that are essentially waiting to be processed, items in xi_eventqueue will be processed and then two tables are updated(before xi tries to actually send the notifications) - xi_events and xi_meta. xi_events keeps record on if an event has been processed or not and xi_meta contains the actual information of the event. Finding notifications events that have not been processed yet can be seen with:
select * from xi_events where status_code = 0 and event_type = 2;


Aside from this, we have identified some issues with the mailing system in Nagios performing slower than expected. We have several markers that indicate that our current system is simply taking too long to process email notifications as well as some not being processed at all. We are currently working on ways we can fix these issues.

In the meantime, truncating these tables might be an option to consider -> xi_meta, xi_events, xi_eventqueue

Thanks!
Benjamin Rowan
Nagios Developer

I am also curious how large your system is, more or less how many hosts/services you have.

Often times even though system metrics look healthy, the size of a system can significantly impact how quickly these notifications are processed. This information can help us determine if there is a significant bottleneck in performance relative to how many possible notifications are possible. Which seems to vary, Thanks!

Benjamin Rowan
Nagios Developer