avoid ipv6 loopback address reported as src or dst of span #1021
Conversation
|
|
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1021 +/- ##
==========================================
+ Coverage 79.06% 81.64% +2.58%
==========================================
Files 137 137
Lines 10923 10923
==========================================
+ Hits 8636 8918 +282
+ Misses 1761 1497 -264
+ Partials 526 508 -18
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
esara
force-pushed
the
ipv6
branch
5 times, most recently
from
e30ea1f to
ee3f85f
Compare
esara
changed the title
pkg/internal/ebpf/common/common.go
Outdated
| @@ -231,10 +231,21 @@ func cstr(chars []uint8) string { | |||
| } | |||
|
|
|||
| func (connInfo *BPFConnInfo) reqHostInfo() (source, target string) { | |||
| oneAddress := [16]uint8{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1} | |||
There was a problem hiding this comment.
This is technically the IPv6 Loopback address. I'm still concerned that we have an event with it, but I don't mind putting a check to verify it's valid.
The Go library code does it like this:
func (ip IP) IsLoopback() bool {
if ip4 := ip.To4(); ip4 != nil {
return ip4[0] == 127
}
return ip.Equal(IPv6loopback)
}
So I think all we need to do is:
if src.Equal(net.IPv6loopback) {
src = ""
}
if dst.Equal(net.IPv6loopback) {
dst = ""
}
There was a problem hiding this comment.
that is much cleaner, updated ....
on another note, does it make sense to report ipv4 loopback addresses in traces?
There was a problem hiding this comment.
I'd say it might have sense. E.g. if you run Beyla on a baremetal host that runs multiple services.
You can always use filters to remove local traces: https://grafana.com/docs/beyla/latest/configure/options/#filter-metrics-and-traces-by-attribute-values
esara
changed the title
address first part of #895