thold plugin 0.4.6 screenshots

[TheWitness]

Also another major issue found that template Filtering not working when selecting Status either Breached/Triggered..Till 0.4.4 it was fine.

I am unable to reproduce this. The Status filters work on both Chrome and Firefox.

[TheWitness]

Commit #1827 should resolve the first issue. I can not reproduce the filter error.

TheWitness

[sukanta]

I have tried with two Different OS (Solaris & CentOS5.6) with fresh install also but same problem, Filter not working. It is showing all the triggered value whatever filter applied.

Also another thing notices that for thold to trigger it is happening as soon as current value match threshold value.For old versions it was grate than threshold value.

Edit:Same problem even in thold 0.4.7

thold.png (52.52 KiB) Viewed 3311 times

[uhtred]

i can confirm the error for status 'Breached' or 'Triggered' and a template filter- with firefox & chrome
result is: selection for Host will change, but u will see all thresholds of level notice and warning ( from every Template)

the problem is the sql request.
<<BREACHED>>
In thold_graph.php line 338 has to be changed from:

Code: Select all

if($_REQUEST['triggered'] == '1') { $sql_where = "WHERE thold_data.thold_alert     !=0 OR thold_data.bl_alert>0"; } /* breached */

to

Code: Select all

if($_REQUEST['triggered'] == '1') { $sql_where = "WHERE (thold_data.thold_alert     !=0 OR thold_data.bl_alert>0)"; } /* breached */

<<TRIGGERED>>
In thold_graph.php line 339 has to be changed from:

Code: Select all

if($_REQUEST['triggered'] == '3') { $sql_where = "WHERE (thold_data.thold_aler     t!=0 AND thold_data.thold_fail_count >= thold_data.thold_fail_trigger) OR (thold_data.bl_alert>     0 AND thold_data.bl_fail_count >= thold_data.bl_fail_trigger)"; } /* triggered */

to

Code: Select all

if($_REQUEST['triggered'] == '3') { $sql_where = "WHERE ((thold_data.thold_aler     t!=0 AND thold_data.thold_fail_count >= thold_data.thold_fail_trigger) OR (thold_data.bl_alert>     0 AND thold_data.bl_fail_count >= thold_data.bl_fail_trigger))"; } /* triggered */

[uhtred]

another problem: Restoral doesn't work

Update:
it looks like it is a problem, if u:
1. create a threshold to a graph with wrong values (High/Low)
2. add a threshold to a graph
--> notice event is triggered
3. u change values of thold template to correct one within the Min Trigger Duration time
--> no restoral event
--> threshold will allways get the status 'Breached'

[TheWitness]

another problem: Restoral doesn't work

Ok, I fixed the one and the second, I'll research later today.

TheWitness

[sukanta]

My All problem resolved now except following:

THOLD: Poller[0] WARNING: Unsupported Field '|QUERY_IFHIGHSPEED|'

Still error coming.

[TheWitness]

Did you update the XML file and requery your hosts? They have to support ifHighSpeed first. If they don't, there is no logic presently to back off to ifSpeed. Maybe we should do that automatically. What do you think?

[sukanta]

Did you update the XML file and requery your hosts? They have to support ifHighSpeed first. If they don't, there is no logic presently to back off to ifSpeed. Maybe we should do that automatically. What do you think?

Yes.Already updated both rrd.php & Interface.xml & 10G link graph starts coming fine in 0.8.7h Cacti but thold is giving error.

[sukanta]

There is Major Performance issue for Large Site:

Without Thold:

10/12/2011 09:50:23 AM - SYSTEM STATS: Time:21.7758 Method:spine Processes:4 Threads:15 Hosts:2754 HostsPerProcess:689 DataSources:40639 RRDsProcessed:18279

With Thold 0.4.4:

0/12/2011 12:20:26 PM - SYSTEM STATS: Time:24.1262 Method:spine Processes:4 Threads:15 Hosts:2754 HostsPerProcess:689 DataSources:40639 RRDsProcessed:18289
10/12/2011 12:16:28 PM - SYSTEM THOLD STATS: Time:57.5140 Tholds:18113 Hosts:0

With Thold 0.4.7:

10/12/2011 10:24:03 AM - SYSTEM THOLD STATS: Time:66.5335 Tholds:18111 Hosts:0
10/12/2011 10:22:57 AM - SYSTEM STATS: Time:175.0682 Method:spine Processes:4 Threads:15 Hosts:2754 HostsPerProcess:689 DataSources:40639 RRDsProcessed:18287

[uhtred]

There is Major Performance issue for Large Site:

what do u mean with that?

[sukanta]

There is Major Performance issue for Large Site:

what do u mean with that?

It is self explanatory:

Without Thold Total time taken by pooling: 21.7758 Seconds

With Thold 0.4.4 Total time taken by pooling: 140 Seconds

With Thold 0.4.7 Total time taken by pooling: 243 Seconds

[uhtred]

i cannot confirm/reproduce that.
SYSTEM THOLD STATS: Time:0.5092 Tholds:203 Hosts:0
look my signature

[TheWitness]

Are you using either RPN Expressions or Baseline alarms in 0.4.7? Please let us know.

[TheWitness]

Turn on your slow query log at say 2 seconds and run the system for a few hours and post your slow query log to a PM or a bug report.