On 21/01/2016 18:23:07, Eduardo Schoedler <listas@esds.com.br> wrote:
I tested a lot the script to remove spikes, but I can't understand how it works.
# scripts/removespikes.pl -l 0.06 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 0.06
No peaks found.!
# scripts/removespikes.pl -l 6 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 6
No peaks found.!
# scripts/removespikes.pl -l 60 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 60
No peaks found.!
# scripts/removespikes.pl -l 0.06 -t 100000000 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 0.06
Max Value set to 1e+08, disabling bin-based chopping
No peaks found.!
# scripts/removespikes.pl -l 10 -t 100000000 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 10
Max Value set to 1e+08, disabling bin-based chopping
No peaks found.!
# scripts/removespikes.pl -l 20 -t 100000000 -v /tmp/port-512.rrd
Running in VERBOSE mode
Limit for bin-based chopping set to 20
Max Value set to 1e+08, disabling bin-based chopping
No peaks found.!
2016-01-21 16:09 GMT-02:00 Adam Armstrong <adama@memetic.org>:
These spikes shouldn't really happen, but sometimes it's hard to weed them out.There might be some code we can use from the ports poller to try to filter these out in future, if i get a chance to take a look.adam.On 21/01/2016 17:23:14, Eduardo Schoedler <listas@esds.com.br> wrote:
Adam,Yes, there is a large value in the table:mysql> SELECT * FROM bill_data WHERE bill_id = 10 AND timestamp BETWEEN '2016-01-20 16:00:00' AND '2016-01-20 18:00:00';+---------+---------------------+--------+---------------+-----------+---------------+| bill_id | timestamp | period | delta | in_delta | out_delta |+---------+---------------------+--------+---------------+-----------+---------------+| 10 | 2016-01-20 16:00:00 | 300 | 420089822 | 353359170 | 66730652 || 10 | 2016-01-20 16:05:00 | 300 | 494265345 | 361777810 | 132487535 || 10 | 2016-01-20 16:10:00 | 300 | 528145357 | 372315740 | 155829617 || 10 | 2016-01-20 16:15:00 | 300 | 346064669 | 260674031 | 85390638 || 10 | 2016-01-20 16:20:00 | 300 | 373899963 | 286552744 | 87347219 || 10 | 2016-01-20 16:25:00 | 300 | 406708364 | 310406142 | 96302222 || 10 | 2016-01-20 16:30:00 | 300 | 353585558 | 286064759 | 67520799 || 10 | 2016-01-20 16:35:00 | 300 | 376104759 | 293666507 | 82438252 || 10 | 2016-01-20 16:40:00 | 300 | 466089158 | 328439085 | 137650073 || 10 | 2016-01-20 16:45:00 | 300 | 1231515051765 | 264605699 | 1231250446066 || 10 | 2016-01-20 16:50:00 | 300 | 321868504 | 244902936 | 76965568 || 10 | 2016-01-20 16:55:00 | 300 | 396602378 | 295232411 | 101369967 || 10 | 2016-01-20 17:00:00 | 300 | 435646403 | 302257751 | 133388652 || 10 | 2016-01-20 17:05:00 | 300 | 574915176 | 406955741 | 167959435 || 10 | 2016-01-20 17:10:00 | 300 | 420821268 | 281749562 | 139071706 || 10 | 2016-01-20 17:15:00 | 300 | 497632123 | 364700162 | 132931961 || 10 | 2016-01-20 17:20:00 | 300 | 491631241 | 374396028 | 117235213 || 10 | 2016-01-20 17:25:00 | 300 | 436369775 | 338883451 | 97486324 || 10 | 2016-01-20 17:30:00 | 300 | 495699422 | 385720549 | 109978873 || 10 | 2016-01-20 17:35:00 | 300 | 456040027 | 366380731 | 89659296 || 10 | 2016-01-20 17:40:00 | 300 | 343876841 | 253132670 | 90744171 || 10 | 2016-01-20 17:45:00 | 300 | 288716438 | 208696427 | 80020011 || 10 | 2016-01-20 17:50:00 | 300 | 249196310 | 200195401 | 49000909 || 10 | 2016-01-20 17:55:00 | 300 | 277595458 | 191080322 | 86515136 || 10 | 2016-01-20 18:00:00 | 300 | 349455603 | 260510508 | 88945095 |+---------+---------------------+--------+---------------+-----------+---------------+25 rows in set (0.00 sec)Thank you!_______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium2016-01-21 3:59 GMT-02:00 Adam Armstrong <adama@memetic.org>:To remove that, go in to the database (phpmyadmin is useful!), look at the bill_data table and delete values which are too large.You shouldn't really get many spikes though!adam.On 21/01/2016 05:39:34, Eduardo Schoedler <listas@esds.com.br> wrote:
Perfect, I'll do that.For the record, here is the "accurate graph":The spike is there too :))_______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium2016-01-21 3:35 GMT-02:00 Adam Armstrong <adama@memetic.org>:Hmm. It actually seems that that graph is returning the billing-data 95th, rather than the rrdtool 95th. That's fine, then.The "quick graphs" are generated from the rrd files of the ports included in the bill. the "accurate graphs" are generated from the billing data held in the database.You should take the value from the "bill summary" box on the billing page though.adam.On 21/01/2016 05:31:03, Eduardo Schoedler <listas@esds.com.br> wrote:
Sorry... 95th percentile._______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium2016-01-21 3:30 GMT-02:00 Eduardo Schoedler <listas@esds.com.br>:Adam,We do 97th percentile billing using this graph, with the value of the legend.Do you have sure this spike is not going to interfere in the calculations?Thank you.2016-01-21 3:19 GMT-02:00 Adam Armstrong <adama@memetic.org>:Just ignore them. If they're not present in the non-rrd billing data, they're irrelevant.To remove ones which don't disappear with default settings, you have to spend half of your life arsing around with the removespikes.pl settings, which is a pain.adam.On 21/01/2016 05:08:52, Eduardo Schoedler <listas@esds.com.br> wrote:
I ran the script for all rrds, but the spikes are still there.How can I remove?Thanks._______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium2016-01-21 3:03 GMT-02:00 Adam Armstrong <adama@memetic.org>:That graph actually isn't billing data, it's just rrd.The "detailed" graph is built from billing data.adam._______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observiumOn 21/01/2016 04:35:58, Eduardo Schoedler <listas@esds.com.br> wrote:
Hi,Forget it, I found something in the site... :)for i in rrd/*/*.rrd; do scripts/removespikes.pl $i; done;Thanks.2016-01-20 17:13 GMT-02:00 Eduardo Schoedler <listas@esds.com.br>:Hi,Is there a way to remove spikes from the billing?Thank you.Regards,--Eduardo Schoedler--Eduardo Schoedler
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
--Eduardo Schoedler
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
--Eduardo Schoedler--Eduardo Schoedler
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
--Eduardo Schoedler
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
--Eduardo Schoedler
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Eduardo Schoedler