How to diagnose what is using all my free tier data

Hi, (one of) my account has been suspended because it exceeds the 100GB free tier.

I was looking at evaluating its cost and I followed instructions at New Relic One Free Tier - FAQ to see what is the current usage and I found an unexpected high usage in the plaform.

How can I drill down into that table? I see the majority of usage is categorized under “Metrics” but 117GB per day is a huge amount of data. I have another application which fits under the free tier that perform more traffic that this one and it consumes about 0.80 GB per day in “Metrics” category.

The only difference from these 2 accounts is that in the second one I’ve added K8S newrelic integration with pixie, is it possible that K8S integration uses all those data?


Hi there @fabn -

Welcome to the community! We have a number of queries in our documentation that you can use or modify to drill down into the details of where your data is coming from. Try these:

Let me know if that does not get you the level of detail you are looking for.

I was able to gather some data using this query:

FROM Metric SELECT bytecountestimate()/10e8 as 'GB Estimate' 
  SINCE '2021-12-01' UNTIL '2021-12-15' 
  FACET metricName LIMIT 20 TIMESERIES 1 day

Now I see this results and as far as I can understand here’s the culprit

If I understand the query just these metrics sums up to about 60GB of data. Now I need to understand if it’s normal that all those data are ingested by k8s integration of if I have something wrong in my cluster.

Is it normal that a relatively small cluster (about ~150 pods running) produces all those data? If yes I think I’ll have to deactivate the integration because it’s an expense we cannot afford right now.

Excellent work @fabn -

Glad you were able to identify the issue. I can’t speak to what’s “normal,” but you may want to take a look at the YML file for the integration and see if there are variables to adjust there, like frequency of data collection, that may limit the data ingest. You could play with some settings to see what gives you actionable insights without putting you over budget.

I upgraded nri-bundle using –set global.lowDataMode=true and data usage dramatically decreased, I’ll monitor in the next few days if it feasible.

Thanks for help.