New Relic Route Change decreased dramatically

In my company we use NewRelic for platform monitoring.

What I have noticed that between 9th and 10th of November, the avg number of route changes (browser interactions) dropped from 160k/day to 20k/day. it stays at circa 20k/day till now. I haven’t changed the application code or the infrastructure it runs on. The number of user logins hasn’t changed as well until now.

Has there been recently any upgrade to javascript newrelic agent or to the algorithm how RouteChange is calculated?

Is there any other reason why it could drop significantly within a single day?

@pbozowski Thanks for the question here! There haven’t been any changes that I’m aware of that would change anything in terms of route changes. Here’s a link to our release notes.

Usually when I see a drop in PageView or BrowserInteraction data, I would look at any recent deployments as I’ve seen many cases where the browser agent was inadvertently uninstrumented on certain application pages. One way to check that the agent is running on every application page is by running the command newrelic.info in the browser console of the app page. If it returns an undefined message, there is no agent running. If it returns a json object of metadata, you know that the agent is properly working.

Were there any changes made to data retention? Were there any drop rules implemented?

Additionally, if you include a link to your application, I’m happy to take a look. Thank you again!

@cfrankenfield Many thanks for your response.

Application links where I observed decrease in browser interactions (for some reason the forum doesn’t allow me to publish links so I replaced dots with underscore):

• emea3_mrtedtalentlink_com
• emea5_lumessetalentlink_com
• emea8_mrtedtalentlink_com
• emea9_sabatalentlink_com

The data retention period was reduced to 6 months in December 2020. I am not aware of any drop rules. I will check if the agent works properly on application pages and update this thread.

I really appreciate your support on this!

@pbozowski I’m happy to help! I meant to also include this link here for troubleshooting installation that I mentioned previously.

Is there another email address you use for your account? I am trying to take a look at your account; would you please be able to provide a link or the account id? Thank you!

Hi @cfrankenfield, apologies for delay in response. My account belongs to Saba Inc (Lumesse Ltd). Please check, it should be attached now. I can send you also my license key if necessary.

What I observed is that our Content Security Policy doesn’t allow to load the script bam-cell_nr-data_net. I haven’t seen this error in the past, maybe this is the root-cause. I will update our CSP and then re-check next week.

I would appreciate any further guidelines after reviewing the data from my company account.

@pbozowski Thanks for the update! I was able to find a browser application on your account called TalentLink EMEA3. Is that the correct one? If not, can you please post a link here to the application?

I ran the query SELECT count(*) FROM BrowserInteraction WHERE appName = 'TalentLink EMEA3' since 60 days ago TIMESERIES and seeing over 100k average and a similar week over week pattern since November with no significant drop. Are you able to post a photo or query that you are using to compare?

And yes, the bam-cell_nr-data_net browser endpoint is fairly new so you may need to add that to CSP. Please keep us posted on the status and you can find more information here in our documentation.

Thank you!

@cfrankenfield I really appreciate your support on this.

Yes, TalentLink EMEA3, TalentLink EMEA5 and TalentLink EMEA8 are applications where I observed the drops between end of October and beginning of November 2020. I attached the charts for reference.

I will of course review the documentation and update CSP to include “bam-cell” script.

emea3.png1558×794 37.8 KB

emea5.png1550×806 38.2 KB

@pbozowski Thanks for the visuals! It looks like I can see a drop around Nov. 8. Were there any deployments or changes made around that time that may have inadvertently affected the install of the browser agent?

Screen Shot 2021-01-15 at 3.53.59 PM.png2054×792 94.4 KB

It looks like you have installed via Java APM agent? I checked for deployment markers though couldn’t find any in the UI. Were there any changes there?

I checked your data retention and that looks fine. I think making sure that all pages are running the agent that are supposed to be is the next best step to look.

Hi @cfrankenfield.

It seems the update of CSP policy solved the problem. I see increase in browser interactions.

We will prepare monitoring for similar CSP violations. It will allow us to tackle similar problems much faster.

Once again, many thanks for your support on this, much appreciated!

Thanks for letting us know @pbozowski and glad to hear the issue has been resolved.