Open ringerc opened 11 months ago
Discussed with DD support. The support team says that "availability-zone" is just the default tag used in the host infra view, and is only used for AWS anyway. So they don't consider sending zone
on metrics vs the UI expecting availability-zone
in the host inventory to be an issue.
But the code is still inconsistent with the docs it links to and the UI's expectations.
I took a quick look at the DD agent code but couldn't immediately see which it sends. I did notice that in my production account the agent isn't sending any cloud AZ info at all for my Azure hosted nodes; there's no cloud metadata attached to the metrics at all, only kube metadata.
Has there been any resolution as to how to set the availability-zone in the infrastructure map view? I am currently passing the zone under the resource attribute cloud.availability_zone
, cloud.availability-zone
, and cloud.zone
. If I just have cloud.availability_zone
it gets mapped to the tag zone
but not availability
From my last conversations with DD folks before I dropped my use of the product there was no fix for this.
https://github.com/DataDog/opentelemetry-mapping-go/blob/5432be02915e24c7e8a294cd4a17877016f9b73d/pkg/otlp/attributes/attributes.go#L53
This mapping sends
cloud.availability_zone
as the datadog tagzone
. But the docs it links to, https://www.datadoghq.com/blog/tagging-best-practices/, say that the tag name should beavailability-zone
.One of them is presumably wrong.
The host inventory uses
availability-zone
so unless the metrics backend on DD useszone
instead for some reason, this code might be mistaken/outdated.