Hi,
I want to track internal web requests pass/fail % and alert on them if the fail % reaches certain threshold. I have tried to use the "Web Requests" dashlet where it shows non entry points but it does not look like the numbers are correct (from pass/fail perspective). Also if i want to filter this dashlet for a particular internal webrequest, , it is filtering the actual web-request which is calling this internal web-request. Another downside is, i don't have a measure to alert on it or to track it.
So i have created measures under "Functional Health" -> HTTP Response code (That measure seems to be BT Evaluation//Filter/Splitting).
For example - For a particular uri pattern - i have created http 200s and http 4xx/5xx. I am planning to use the rate calculation to get the error rate. i would like to do this for most of the uri patterns. is this the efficient way to do it? One other thing i observed is, as soon as i create a measure under "Functional Health" > HTTP Response code, that measure is coming up under "Diagnose Events" - > Errors. can we avoid this? Even though i created http 200s, it is coming up in that errors dashboard.
Answer by Venkata T. ·
Andi - i have opened this RFE for this - RFE - Better support for Services tracking
Answer by Andreas G. ·
Hi
I took a look at your session you sent me - sorry for that taking that long.
If I understand you correctly you basically want to see the "Failure %" for your individual URLs just in the same way as the "Web Requests Dashlet" shows it to you when you change it to the mode "Show -> All". There you can see the Failure % per URL and not Per PP - is that correct?
Unfortunately there is no real way to get this as a measure. Not as a regular measure nor as a BT. The only option right now is to look at the data through the Web Request Dashlet. Which is most likely not what you are looking for as I assume you want to chart this over time to see trends and get alerted through Incidents.
I think what you are explaining is a great RFE - but - should be extended to say something like "Better support for Services". Because this should not end with Failure Rate but also Response Time, Baselining, ...
Andi
Answer by Andreas G. ·
Hi. I will need to discuss this internally on whether there is an option for that scenario. We two have also been in email exchange. If you could - please share a couple of PurePaths and either attach them to this forum thread or send them via email.
Answer by Andreas G. ·
Hi
All Error Detection Rules will show up in the Errors Dashlet. Even though you define a Rule for HTTP 200. So - thats nothing you can change,
To your particular scenario. There is an out-of-the-box Error Rule that detects "HTTP 4xx Inter-Tier Responses". It will mark that PurePath with an error. If you want that the PurePath is actually marked as failed you can change the Impact setting to "Mark Transaction as Failed". This will then mark the whole PurePath as failed and the PurePath will then also contribute to the pass/fail %.
When opening the Errors DAshlet you should now also see the Internal Web Request URLs in the bottom section of the Errors Dashlet that had the HTTP 4xx.
Let me know if this makes sense - and if this helps.
Andi
Andi,
We don't want the parent call (to fail as the parent call is not really failing for the end-user. In our case, the price info is not available. So we are looking at tracking only the internal webrequests failure % by each uri. We also want to chart them for any given uri based on the requirement. we are planning to do this for our webserver calls and 3rd party calls.
So i was thinking of doing using BT. So i have created a BT and tried to split it by a method argument which actually has these webservice/3rd party calls as an argument. Right now, it is still not working. Even if it works, in that somehow i need to include the status for these calls.
JANUARY 15, 3:00 PM GMT / 10:00 AM ET