long time gap to receive the alert email from cacti

[saya]

Hello all,
I applied thold to the devices and was able to receive the alert email informing which device was down. But I noticed that there's a quite long time gap(some could be up to 11mins) from the time cacti sent an email to me to the time I received the email...May I ask what caused the delay in sending the alert message?

Thanks a lot!

Saya

[Howie]

Hello all,
I applied thold to the devices and was able to receive the alert email informing which device was down. But I noticed that there's a quite long time gap(some could be up to 11mins) from the time cacti sent an email to me to the time I received the email...May I ask what caused the delay in sending the alert message?

Thanks a lot!

Saya

Have you checked where the delay occurred? The mail headers will tell you when the first MTA in the chain (the Cacti server's mail server) sent the message.

It could be a delay in your mail system rather than in Cacti/Thold sending mail...

Also, if you have set your poller ping_failure_count to more than 1 (perhaps 2 here?) then it will take that many poller cycles to detect a down host.

[saya]

hello Howie
I'm not sure where the problem is...
The email is only saying "Host Notice : DEVICE-NAME(device-ip) returned from DOWN state"...
If the delay is caused by my main system, does it mean I dun't have any method to cut down the delay?

[Howie]

hello Howie
I'm not sure where the problem is...
The email is only saying "Host Notice : DEVICE-NAME(device-ip) returned from DOWN state"...
If the delay is caused by my main system, does it mean I dun't have any method to cut down the delay?

Well, try some of the checks I suggested

That way you at least would know if it's your mail system or Cacti that is delaying. And then, if it is Cacti, you can look at the Poller settings to see if you have Failure Count set to higher that 1.

[saya]

Hello Howie, thank you a lot for your reply...
I monitored my device and drawed a graph on the up/down situation and time when receives email from cacti..

for the above chart, the failure count is set to 3. You can see the time gap for sending and receiving email as my above two post said is not so obvious...but there is quite long gap from time a switch got down/up to time cacti sent this down/up alert to me...especially the one at 10:49am in red(around 9 min)..the fastest reply could be within 1 min like the one at 12:11pm(bottom chart) which was less than 1 min after it returned back from down state..
for the bottom chart, both failure count and recovery count were set to 1. (I wonder why there's no recovering state but up state in stead..)And unluckily, two down alerts were missed(marked in red where they should have be)...

So my question is:
1, what might cause the slow reply(from the time a device triggered from up tp down to the time cacti send this event to me)?set failure count is only like to make cacti refresh device status more quickly but not shorten the gap actually..
2, what might cause the loss of alerts?

Thank you for ur reply!!

Regards,
Saya

[saya]

Hello, I noticed that a sms alert always coming earlier than an email alert..can any one tell me why?
and what may cause the loss of some alert?

Thank you!!

[loinguyen]

Hello Howie, thank you a lot for your reply...
I monitored my device and drawed a graph on the up/down situation and time when receives email from cacti..

for the above chart, the failure count is set to 3. You can see the time gap for sending and receiving email as my above two post said is not so obvious...but there is quite long gap from time a switch got down/up to time cacti sent this down/up alert to me...especially the one at 10:49am in red(around 9 min)..the fastest reply could be within 1 min like the one at 12:11pm(bottom chart) which was less than 1 min after it returned back from down state..
for the bottom chart, both failure count and recovery count were set to 1. (I wonder why there's no recovering state but up state in stead..)And unluckily, two down alerts were missed(marked in red where they should have be)...

So my question is:
1, what might cause the slow reply(from the time a device triggered from up tp down to the time cacti send this event to me)?set failure count is only like to make cacti refresh device status more quickly but not shorten the gap actually..
2, what might cause the loss of alerts?

Thank you for ur reply!!

Regards,
Saya

Have you fixed that issue, I am also having the same issue as your post. I have posted on Cacti forum that no one has only me this problem. If you have fixed your error please help me.

[fahrizal]

Me too, have problem same please share solution if this problem solved