Loosen domain name scrubbing in span.domain tag

[gggritso]

Problem Statement

Relay scrubs domain addresses like my.app.service.com to something like *.service.com. This is okay for a lot of cases, but for some users, subdomains are meaningful to more than just 2 segments. This scrubbing causes different domains to be mashed together in the Requests view.

Solution Brainstorm

• allow more segments (e.g., leave the last three segments, or more)
• no scrubbing at all (surely we can't allow that, there are too many unique values)
• customizable domain scrubbing (let people to configure how domains are scrubbed)
• allow self-hosted users to turn domain scrubbing off (people who have their own infra might want to turn this off, and bear the weight of high cardinality)

[jjbayer]

Domain scrubbing is definitely arbitrary right now, because the second level domain might be just as high-cardinality as the subdomain.

@gggritso what if we approach this in a data-driven way, i.e. collect a sample of unscrubbed domains from from the spans dataset and see what kind of scrubbing would yield the best results? One thing we could immediately evaluate is subdomain removal vs. scrubbing of integer / hex IDs. Is this something you'd be willing to drive?

[aldy505]

I think I'd prefer this one:

allow self-hosted users to turn domain scrubbing off (people who have their own infra might want to turn this off, and bear the weight of high cardinality)

It doesn't disturbs the behavior on SaaS and it has a toggle for it.

[gggritso]

@jjbayer that approach makes sense! I'd be happy to do that at some point, though not sure when I'll have a chance

[aldy505]

Since the implementation on #3813 also works for this issue, we should close it.