How to setup newrelic network monitoring in SCCOS

HaukeSt

Hi Sprykees,

We have a project based on SCCOS that has not gone live, yet.

We are facing some performance issues with a third party system and it would be great to confirm that the issue is not due to some network issue.

Has someone already setup the network monitoring (with newrelic) and could help with that?

Newrelic is already reporting on transactions and db queries just fine.

Thanks

Nataliia Potapova

Hey @HaukeSt ,

I'm not sure 100%, but can this course help you?

thomas.lehner

Hello @HaukeSt

If you NewRelic APM is already reporting transactions, can you please check whether "External Services" is already populated?

If it is not, it would be worthwhile to doublecheck whether your APM is setup in the way as described here.

It should list endpoints external to your application, showing you their response times. Please note that this will not be the best way to diagnose true network problems, as you will still have the applications on both sides in the mix.

Depending on how the systems are connected with one another and whether your remote system will respond to a ping, you could ping the endpoint and analyse the returns by setting up a Jenkins job that does the pinging (replacing google.com with your IP/DNS name).

You can then either pipe the output to an S3 bucket for further analysis, or if the issues show frequently, you can also just watch the console output of this Jenkins job.

Side note: You will not normally be able to ping Infrastructure Components on PaaS side. They do not respond to ICMP.

If you are connected via Site2Site VPN your gateway monitoring might also be able to shed some light on connection stability and performance, as well.

HaukeSt

Hello @Nataliia Potapova, Hello @thomas.lehner ,

Thank you for sharing the link to the course about performance monitoring and the link to the documentation respectively :).

We followed the instructions on setting up the newrelic monitoring for the cloud and I do see some data there but it seems something is missing because I do not see any breakdowns by host (maybe because we only have one instance running for each application?). And distributed tracing seems not to be enabled. (There is no data in the corresponding tab for Glue or Zed (In this project we do not use Yves)).

For Glue the External Services are populated. For Zed they are not.

I would really like to see times taken by Glue and Zed respectively + the times spent on the network in between if that is possible.

I tried to use traceroute using the jenkins job but a lot of hops were not returning anything and the job was at some point deleted. (I don't know by whom.) And I will try using ping, too.

thomas.lehner

@HaukeSt

While might not be able to directly support with the NewRelic setup there is a chance that we can help. Can you please create a Question Case for it and i will see that it reaches the right people to evaluate whether and how we can best assist in this case.

Regarding the Jenkins Jobs: Jobs you create manually (and any UI configuration for that matter) in Jenkins will not survive a re-provisioning of Jenkins - which can happen during deployment or recovery after a crash. They are ephemeral. To persist them, you would need to configure them in your project code like so .

HaukeSt

@thomas.lehner Ok, I or one of my colleagues will create a question case.

Thanks for the support!

fsmeier

Heyhey,
jfyi: I moved this thread into the "Spryker Dev Environment" to keep the main category clear for posts from Spryker itself :)

All the best,

Florian