[pkg/ottl] Accessors for profile attributes #39681
Conversation
rockdaboot
requested review from
TylerHelmuth,
bogdandrutu,
evan-bradley,
edmocosta and
a team
as code owners
rockdaboot
force-pushed
the
profiles-attribute-accessors
branch
from
3d6044b to
ecc282f
Compare
rockdaboot
changed the title
rockdaboot
force-pushed
the
profiles-attribute-accessors
branch
from
ecc282f to
6745566
Compare
| } | ||
| tCtx.GetProfile().AttributeIndices().FromRaw([]int32{}) | ||
| for k, v := range m.All() { | ||
| if err := PutAttribute(tCtx.GetProfile().AttributeTable(), tCtx.GetProfile(), k, v); err != nil { |
There was a problem hiding this comment.
I'm probably missing something, but I wish we could make it more efficient. I understand why PutAttribute/ AddAttribute are that complex, but considering all the complexity around the profiles attributes read/write operations, and how slower it probably is compared to other pdata attributes (specially inside a loop), IMO, all the space complexity/reusability benefits we got from using this structure for profiles, maybe didn't worth it. It would be nice to have some pdata helpers/wrappers that makes it faster/easier to handle (just a thought, not a suggestion).
There was a problem hiding this comment.
I basically agree, but would wait until it turns out to be a bottleneck in the real world. Though, if we already know that there are use cases were hundreds or maybe thousands of attributes exist per message, then we should bring it up at the profiling SIG meeting.
|
Please mark it as ready for review when open-telemetry/opentelemetry-collector#12798 gets merged. Thanks. |
rockdaboot
force-pushed
the
profiles-attribute-accessors
branch
2 times, most recently
from
3657b38 to
df8e707
Compare
rockdaboot
force-pushed
the
profiles-attribute-accessors
branch
from
df8e707 to
3b1d852
Compare
github-actions
bot
requested review from
braydonk,
carsonip,
ChrsMark,
crobert-1,
dineshg13,
dmitryax,
fchikwekwe,
greatestusername,
IbraheemA,
jade-guiton-dd,
JaredTan95,
lahsivjar,
liustanley,
mackjmr,
MikeGoldsmith,
mwear,
mx-psi,
portertech,
shalper2,
songy23 and
tigrannajaryan
rockdaboot
force-pushed
the
profiles-attribute-accessors
branch
from
53d4312 to
cbd181f
Compare
|
@edmocosta The pdata PutAttribute function now has been merged and adopted in this PR. I'd still wait for #39952 until marking this PR as ready to review. |
<!--Ex. Fixing a bug - Describe the bug and how this fixes the issue. Ex. Adding a feature - Explain what this achieves.--> #### Description Access to these lookup tables isn't really useful, as they only provide a slice of values, and we can't check which one is being used by the current profile. Also, with open-telemetry/opentelemetry-collector#13075, the lookup tables are moving out of the profile and into a new dictionary object. So as a first step to the proto migration, this removes access to the lookup tables for a profile. The replacement for this is #39681, which will give acces to the profile attributes, as we do with other signals and abstract away the lookup tables. --------- Co-authored-by: Edmo Vamerlatti Costa <[email protected]>
Description
Adds support for
profile.attributesto OTTL.This allows users to use profile.attributes without deeper knowledge of the protocol internals.
Currently, users have to deal with attribute_table and attribute_indices, which is error-prone and fragile when the experimental profiling signal changes internal representation.
This change depends on open-telemetry/opentelemetry-collector#12798 and thus includes that PR. It will be removed as soon as possible.
The profiles support for the transformprocessor currently relies/waits on this change.