-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Mark OpenTelemetry as ready for production use #24672
Comments
@AlexanderEllis - we have some customers that are interested in using this plugin, but are hesitant to do so until this issue is addressed. Do you have any input regarding whether there is any further work needed on this plugin before the status can be changed? Additionally, do you have any other thoughts regarding whether we can change the status of this plugin? Thanks! |
cc @htuch as extension owner |
Looking at https://github.com/envoyproxy/envoy/blob/main/EXTENSION_POLICY.md#extension-stability-and-security-posture, I think the main open questions from me would be:
@yanavlasov @krajshiva I think we eventually do want to have this as stable / production-ready for our own use cases (but nothing pressing today). |
I'll try to take a shot at these to keep this from falling off the radar....
|
OK, @yanavlasov do you think we can get someone from EPT to help with working through 2/3 here? I'm happy to review any PRs. |
Sorry for the delayed response here — holidays and travel kept me away from GH for a bit. @ashishb-solo I believe the functionality is all set for production use cases — the only potential improvement I'm aware of is a nice-to-have (supporting either HTTP exporting or the existing gRPC exporting). Maybe @K-Prabha can weigh in here re: use in production? @htuch I don't believe anyone has done a code audit, but that would be great to see (and I would definitely appreciate the additional eyes on it). Similarly, fuzzing for the context extractor would be great to see, as it's doing some work parsing headers. Happy to review PRs as well! |
This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions. |
From this comment on the parent tracking bug, it sounds like @lakamsani may also be using the extension with live traffic. @lakamsani, we're working towards marking the extension for production use, and it would be great to hear how it has done in production so far and any potential issues that have come up (or lack thereof). |
Hi @AlexanderEllis sure. We are in pre-production testing right now. Expect to use it in production from May onwards. |
This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions. |
I have a report that an Envoy configured with 100% sampling will drop spans compared with a similarly configured Lightstep/OpenTracing tracer, the one that was removed several releases ago as we put our focus on the OTel tracer. See #25102 It's not clear what kind of tuning parameters there are for the OTel tracer library, but none are documented for Envoy here: https://www.envoyproxy.io/docs/envoy/latest/api-v3/config/trace/v3/opentelemetry.proto.html |
Hey @jmacd, thanks for bringing up that case, very interesting. I can take a look at adding more documentation on the tracer configuration. It does support two additional runtime settings that it uses for span exporting, |
Hey Alex,
These values also don't work in the envoy config itself. So it's not just
the docs...
Thanks
Dan
…On Fri, Apr 7, 2023 at 5:00 PM Alex Ellis ***@***.***> wrote:
Hey @jmacd <https://github.com/jmacd>, thanks for bringing up that case,
very interesting. I can take a look at adding more documentation on the
tracer configuration. It does support two additional runtime settings that
it uses for span exporting, tracing.opentelemetry.flush_interval_ms and
tracing.opentelemetry.min_flush_spans, but those aren't included in the
autogenerated proto docs you linked 👍
—
Reply to this email directly, view it on GitHub
<#24672 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAD4PPZWORFXR74XGLUAGGTXAB55FANCNFSM6AAAAAATHCJUXA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
@DanTulovsky The plot thickens! Just to confirm for the sake of posterity, these are separate from the configuration settings in the OpenTelemetry config proto (the docs @jmacd linked above), but if you're setting the runtime values and they're not being respected, something must be up. I'm traveling for the next week or so, but once I'm back at my usual computer I can dig in a little more to see why these runtime configurations aren't being read correctly. |
Sorry, I must have been doing something wrong. When I tried the runtime
config envoy was failing with it being invalid. But today it's not. :/
I'll try setting that value and see if it makes a difference.
Thank you.
…On Fri, Apr 7, 2023 at 6:16 PM Alex Ellis ***@***.***> wrote:
@DanTulovsky <https://github.com/DanTulovsky> The plot thickens! Just to
confirm for the sake of posterity, these are separate from the
configuration settings in the OpenTelemetry config proto (the docs @jmacd
<https://github.com/jmacd> linked above), but if you're setting the
runtime values and they're not being respected, something must be up.
I'm traveling for the next week or so, but once I'm back at my usual
computer I can dig in a little more to see why these runtime configurations
aren't being read correctly.
—
Reply to this email directly, view it on GitHub
<#24672 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAD4PP3M32AH4G3ZQN7GQM3XACG25ANCNFSM6AAAAAATHCJUXA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
I can confirm that the setting to increase batch size works as expected.
Would it be possible to expose a setting to control compression?
…On Mon, Apr 10, 2023 at 9:28 AM Dan Tulovsky ***@***.***> wrote:
Sorry, I must have been doing something wrong. When I tried the runtime
config envoy was failing with it being invalid. But today it's not. :/
I'll try setting that value and see if it makes a difference.
Thank you.
On Fri, Apr 7, 2023 at 6:16 PM Alex Ellis ***@***.***>
wrote:
> @DanTulovsky <https://github.com/DanTulovsky> The plot thickens! Just to
> confirm for the sake of posterity, these are separate from the
> configuration settings in the OpenTelemetry config proto (the docs @jmacd
> <https://github.com/jmacd> linked above), but if you're setting the
> runtime values and they're not being respected, something must be up.
>
> I'm traveling for the next week or so, but once I'm back at my usual
> computer I can dig in a little more to see why these runtime configurations
> aren't being read correctly.
>
> —
> Reply to this email directly, view it on GitHub
> <#24672 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AAD4PP3M32AH4G3ZQN7GQM3XACG25ANCNFSM6AAAAAATHCJUXA>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
|
Hi Envoy team! Thanks! |
This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions. |
Title: Mark OpenTelemetry as ready for production use
Description: We would like to transition the OpenTelemetry plugin status from
wip
tostable
and labelsecurity_posture
asrobust_to_untrusted_downstream
. According to the plugin's original author, it is essentially feature complete[1], and only awaiting fixes for reported issues. As it currently stands, this plugin has been available for multiple releases by now and there have not been many major issues reported, so we are interested in encouraging adoption of this plugin and declaring it suitable for public use.Relevant Links:
[1]: #9958 (comment)
The text was updated successfully, but these errors were encountered: