Sean Marciniak

Mmm, this is interesting. I haven't had the chance to look, but is this the default value for gostatsd? Should we add some further clarifying docs on it to help...

@jkowall I see this across all types since "background" tasks have the potential to leak data such as appending the SQL query to a metric capturing latency (not my greatest...

Hey @Oberon00 > What do you mean by prefix? I had made an assumption that the labels keys would be transformed to something like: `user.name --> sensitive.user.name` for example. >...

>Providing some extensions especially for the collector that help to identify PII data and to remove or to anonymise it based on configurations (e.g. "strip the last bytes from net.peer.ip...

> the word "resource" has specific meaning in OTEL, you're overloading it by saying "resource classification". You are classifying attributes, not resources. Both are being classified here, attributes have their...

By the end of the day, I'll publish my mini prototype so others can play with themselves

As I mentioned earlier, here is the prototype I had mentioned https://github.com/MovieStoreGuy/data-classifier, it is a rather stripped down only preserving attributes and resource (without any of the values or nested...

To @jamesmoessis point, classifications are coupled to the application deployment. Meaning that application sitting on the network edge would define `net.peer.ip` as PII however, an internal service behind the firewall...

Hey @yurishkuro , > Are those organizations prepared to commit significant engineering capacity to developing and maintaining this non-trivial capability in every language supported by OTel? I am not sure...

@pirgeo , I think you can get away with this without the need to re-aggregate. You can follow similar principle of MapReduce to get around it, meaning: - Filter the...