0
Under review

Monitoring locations not reliable - building links between them

InteractMarketing 8 years ago updated by Konstantin (Here to help) 7 years ago 11
I have set two locations US - Dallas and CA - Vancouver to monitor my servers by pinging them each 3 minutes. Yesterday (May 15th) both locations returned false errors, when all servers were available to the public.

I hope to see ability to link or create some kind of "if" statements for monitoring locations including local agent as last step in that chain to avoid false alarms.


Something like:


( US-Dallas || CA-Vancouver ) && MyLocalAgent01 THEN error

Thanks, we will look to what's happened with agents.


As a quick fix. Why don't you just replace US-Dallas with MyLocalAgent01 ? I.e. have two locations CA and local one.

Well, yesterday I have got multiple false errors even from CA (between 16:15 EST and 16:27EST). If there will an option to create dependency between monitoring locations and local agent then it can eliminate  those false errors.


Good it is still beta testing :)

False errors may happen with any location due to network-related problems. That's why you have 2nd location. I didn't mean you should specifically remove US-Dallas. Idea was to add your local agent as one of monitoring locations.

Yeah, but idea is to monitor an infrastructure properly and avoid any false errors and I think without relations between monitoring locations (them self) and a local agent it will be not possible to avoid and become frustrating for sys admins.

Well, I have decided to record false errors. So far I have done 10 readings with following errors rates (average per day) which were false (servers/devices were up when Anturis returns errors)


HTTP request:

US - Dallas: 0.21%

CA - Vancouver: 2.41%


PING request:

US - Dallas: 0.04%

CA - Vancouver: 0.16%


JFYI 

thanks to your observations we found and fixed a problem that led to some false errors since 29th of May.

Further observations on daily reports and false errors. Below data are average per day based on 25 readings:


HTTP request:

US - Dallas: 1.28%

CA - Vancouver: 1.18%



PING request:

US - Dallas: 0.11%

CA - Vancouver: 0.16%


As you can see some values decreased, some increased and some stay on same level.

Could you please let me know exactly how you measure those percentages? Thanks a lot!

First I'm comparing daily reports from Anturis vs WebPulse then I go to Anturis site and checking graphs. If both monitoring locations shows real issue - it is not false error (rare) if one of them is OK and another comes as an error I recorded it as false error for that location. Then I'm checking if daily there are more false errors for that location - if are then all precedent of errors from daily report is going towards that location. If not then is split between locations based on occurrence. 


Simply - it is looks pretty bad when daily report shows that server has 3% errors when was up all the time.

Well guys, I have no idea to what market are you aiming but for small US company like us, which host our clients websites those issues above will make decision to do not go for paid version, compareing to "local" competition like WebsitePulse.com 


If you guys want I can send some screenshots where you can see that Anturis is throwing errors or issues when WebsitePulse is not and site or server is available, so there is false error on Anturis side.


But anyway I'm glad that I have found some other bugs which were resolved.

If it is the case that anturis triggers a problem, when everything is fine, then yes, we'd appreciate you point to several such problems. Just post the component name and starting time here. Thanks!