Route console log to new relic logs

[murphpdx]

I'd like a setting in the new relic brower agent that will automatically hook into the console.log/console.error functions and send the logs to new relic logging. I would also like it to not log in the browser. Alternatively, an api call that will send logs to new relic. It does not have to hook into console but it would be nice if it did.

Summary

Forward logs to new relic via the browser agent.

Desired Behavior

Conditionally override the console.log, console.error, console.debug, console.info etc functions so that they send logs to new relic and do not log in the console.

Possible Solution

Additional context

[workato-integration]

https://issues.newrelic.com/browse/NR-168151

[patrickhousley]

Hi @murphpdx and thanks for the feature request. This is actually on our roadmap for the quarter starting in January. We did do some work in a past version of the agent to gauge the data that would be sent if we automatically instrumented the console global. Unfortunately we found that extremely large objects are sometimes logged to the console in production systems resulting in a performance issue when we have to serialize it for sending to New Relic.

If you, or any community member or customer, has any suggestions around what you would like to see with this feature, please let us know either here, in our forums, or by reaching out to customer engagement or support.

Some things that we are trying to consider:

• Do we need to implement log levels, what are they, and what do they map to on the console?
• What limits can/should we put in place on what gets serialized for logging?
  • Object depth limit?
  • Only flat objects?
• Do we need an API to allow customer code to inspect a log before it is serialized?
  • Log filtering
  • Custom serialization
• How/Should we handle obfuscation of log data?

[workato-integration]

https://new-relic.atlassian.net/browse/NR-168151

[sconzof]

Hi! Piggy-backing on the request from @murphpdx
This feature set would be very useful for some of our internal applications.

[CavalcanteLeo]

any news about this?

[patrickhousley]

@CavalcanteLeo this work is on our roadmap to start this quarter.

[FranciscoKloganB]

Just shimming in, this feature might be nice as an opt-in.

Some projects allow the use of console.debug (and friends) freely in the code for development reasons and remove them for production builds using vite/webpack/rollup directives.

[CavalcanteLeo]

For me, instead of sending the console.log automatically, I would prefer something to manually send it to new relic, something like newrelic.log(text: string), and on the frontend I can abstract it, and create a function with the logger I want to use...

const useLog = (text: string): void => {
  newrelic.log(text) // here sends the log to new relic [newrelic.info(), newrelic.error()]
  console.log(text)  // Here I could use some other log, like Pino or Winston. And It will get rid in production as well
}

in this case, we don't rely on the console.logs only.

Most of the frameworks get rid of the console logs when building the prod (which is correct)

[caviyacht]

Optionally allowing each method in a config would be preferred. Our main focus is to get console.error captured as it would help track down a ton of problems.

The console.log could become a ton of noise and high data ingest, but toggling it on and off at will would be super beneficial.

[piyushchauhan2011]

Thank you for good work on logs