cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

This product reached the end of support date on March 31, 2021.

Web Services Monitoring (Response Time and Failures)

wilson_reis1
Inactive

Hi all,

We have an environment that highly
based on SOA. So all most of the application process is split between different
servers using shared web services to handle the requests.

We would like to have a better
visibility and alerts based on those web service
calls.

Basically what we wanted is a way to monitor the data that’s shown from the
web services dashlet, creating alerts based on the web services and response time:

cid:image002.png@01D1AFAA.65533CB0

First of all I had a look into the Web Service
measures that we can create, but there is no measure for failed web services or their response time (even though we have that on
webservices dashlet).

cid:image003.png@01D1AFAA.65533CB0

I tried to work with the business transaction,
filtering by specific web service requests.

However I run into 2 issues which can be seen
in the screenshot below:


  • -The
    failed count only add when the whole purepath returns as a 5xx error, which
    doesn’t mean that the internal web service calls didn’t had errors (Purepath
    can be “yellow” state with 5xx errors on
    internal web service calls and doesn’t add to “failed” measure from the BT
    since the purepath returns as a HTTP 200 at the top of the tree).

  • -Even
    if I add the measure “Http 5xx Errors (Internal)” it not necessary reflect
    errors within the filtered web service. Since the whole structure from is based
    on web service calls (SOA Environment), usually there are more than one web
    service calls per purepath, which means that
    the “Http 5xx Errors (Internal)” will add to count if there are any 5xx errors
    with any of the web services inside purepath filtered to the BT.

So, I would like your ideas
on how would be the best way to set up this
monitoring for the hundreds (I mean it
literally) of internal web service calls.

Besides that,
I’ve looked into the community and found out there are other customers who have
a SOA environment and are very interested on the improvement of dynaTrace
regarding the monitoring of those environments.

I think this RFE would be what we need RFE: Business Transaction for Web Request, but from Andreas Grabner last
reply (1 year ago) it doesn’t seem that it will be available anytime soon.

2 REPLIES 2

dave_mauney
Dynatrace Champion
Dynatrace Champion

I have had similar experience in the past. See this post for more details: https://answers.dynatrace.com/questions/105937/web...

What I did to work around the issue was create one web services time measure for a particular service. Then, using that original measure XML from the system profile as a template, I used the output of the web services dashlet to create the XML needed in the system profile to add the other measures with a Notepad++ macro. You can create a separate system profile for this purpose, import it and then cut and paste the new measures into your actual system profile. This takes a lot of the pain out of the many measures that need to be created.

We gave up on the failure rate part given the lack of measures available. Perhaps you might be able to track specific exceptions or capture faults being returned to track that aspect?

If you need detailed instructions for the measure creation part, send an email to dave.mauney@dynatrace.com and I will dig up the instructions I provided to the customer for this process.

HTH,

dave

wilson_reis1
Inactive

Hi Dave,
I appreciate your help, since this is duplicated topic I'll keep tracking on your.

Thank you,

Wilson