Date   

Re: Jaeger graduation due diligence

Matt Klein
 

Related to comments from @Joe Beda and @William Morgan, my general feeling is that frankly very few organizations get actual value out of distributed tracing currently, no matter what provider they are using (OSS tools, commercial, etc.). I've seen countless orgs put a lot of effort into getting things working only to then find that developers have no idea how to interact with the system, can't find the trace (the prototypical gant chart view) that would help them, etc. The ROI just isn't there in the general case.

With that said, my feeling is that while the current UI/UX around DT is not all that useful, the underlying data is without a doubt incredibly useful. Organizations like Uber have built ancillary tooling that mine the trace data and provide useful services around issue isolation, debugging, etc. (Commercial providers like Omnition and Lightstep are also heading in this direction.)

A system like Jaeger is an amazing base by which to build such tooling in the open source domain, and I think there is already enough usage of it throughout the industry to easily justify graduation and promote the creation of better ancillary tooling within the project itself (as @Yuri Shkuro already mentioned, the creation of OpenTelemetry will ultimately allow Jaeger to focus more on the backend system vs. the client libraries, collectors, etc. which is a great thing).

Thanks,
Matt

On Thu, Oct 10, 2019 at 4:02 PM alexis richardson <alexis@...> wrote:
Super post wm

On Thu, 10 Oct 2019, 15:24 William Morgan, <william@...> wrote:
On the topic of "the state of tracing", we JUST added tracing support to Linkerd. As someone with no real horse in the DT race, but with a desire to actually provide a concrete solution to our users, it was a particular painful exercise in navigating a complex landscape that at times seemed to consist solely of projects that were either deprecated or not ready for production. Writeup in https://linkerd.io/2019/10/07/a-guide-to-distributed-tracing-with-linkerd/.

(But Jaeger is great.)

-William


On Tue, Oct 8, 2019 at 8:03 AM via Lists.Cncf.Io <jbeda=vmware.com@...> wrote:

A couple of thoughts.

 

We should talk about the state of “tracing” in general.  It is useful to say that someone like Facebook is using it but it is also important to call out that they aren’t using Jaeger (right?).  Many many of the orgs doing this are very mature in their tooling and often have their own solutions in place.  Many more mature enterprises aren’t set up yet to really drive tracing.  Agree that OpenTelemetry could change that over time.

 

It is worth talking about the instrumentation/journey that many companies take.  Many start with metrics and logging (both structured and unstructured) long before they are mature enough for tracing.  I don’t kow if the CNCF has any advice for that.

 

Also worth noting that newer architectures that have more structure around user code (“serverless”) provide a great opportunity for automatically making tracing happen.  This will mature as open serverless frameworks become more mature.  (Things like Lambda are a closed system where they only integrate with their own tracing services).

 

Joe

 

From: <cncf-toc@...> on behalf of alexis richardson <alexis@...>
Date: Tuesday, October 8, 2019 at 5:45 AM
To: Yuri Shkuro <ys@...>
Cc: Matt Klein <mattklein123@...>, Alexis Richardson via cncf-toc <cncf-toc@...>, Gary Brown <gbrown@...>
Subject: Re: [cncf-toc] Jaeger graduation due diligence

 

Thanks Yuri and Matt.  This has been helpful for me.

 

I realised what was bothering me.  I think we should graduate Jaeger; but when we do so I think we should also provide some thoughts on the state of adoption.  Users will I hope feel more enlightened if we share that 1. The project is solid and has enough real use, 2. But the user base remains narrow, 3. Which will change as more ms apps are built and scaled, and tools like opentelemetry emerge, 4. And here is how you can help. 

 

 

 

On Mon, 7 Oct 2019, 21:44 Yuri Shkuro, <ys@...> wrote:

Thanks, Matt.

 

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

 

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

 

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

 

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:

Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

 

Weaveworks is not an enterprise, it is a small isv. 

 

Has jaeger found much traction among, say, non members of CNCF? 

 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  

 

 

 

 

On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

 

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.

 

 

On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:

Are you specifically referring to enterprise reselling/platform and not end-users?

 

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:

Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

 

Weaveworks are a happy user, as documented here.  

 

 

On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:

Hi folks,

 

Here is Jaeger's due diligence document for review prior to the graduation vote:

 

Please take a look at this document as well as the graduation PR and comment away!

 

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

 

Thanks,

Matt


Re: Jaeger graduation due diligence

alexis richardson
 

Super post wm


On Thu, 10 Oct 2019, 15:24 William Morgan, <william@...> wrote:
On the topic of "the state of tracing", we JUST added tracing support to Linkerd. As someone with no real horse in the DT race, but with a desire to actually provide a concrete solution to our users, it was a particular painful exercise in navigating a complex landscape that at times seemed to consist solely of projects that were either deprecated or not ready for production. Writeup in https://linkerd.io/2019/10/07/a-guide-to-distributed-tracing-with-linkerd/.

(But Jaeger is great.)

-William


On Tue, Oct 8, 2019 at 8:03 AM via Lists.Cncf.Io <jbeda=vmware.com@...> wrote:

A couple of thoughts.

 

We should talk about the state of “tracing” in general.  It is useful to say that someone like Facebook is using it but it is also important to call out that they aren’t using Jaeger (right?).  Many many of the orgs doing this are very mature in their tooling and often have their own solutions in place.  Many more mature enterprises aren’t set up yet to really drive tracing.  Agree that OpenTelemetry could change that over time.

 

It is worth talking about the instrumentation/journey that many companies take.  Many start with metrics and logging (both structured and unstructured) long before they are mature enough for tracing.  I don’t kow if the CNCF has any advice for that.

 

Also worth noting that newer architectures that have more structure around user code (“serverless”) provide a great opportunity for automatically making tracing happen.  This will mature as open serverless frameworks become more mature.  (Things like Lambda are a closed system where they only integrate with their own tracing services).

 

Joe

 

From: <cncf-toc@...> on behalf of alexis richardson <alexis@...>
Date: Tuesday, October 8, 2019 at 5:45 AM
To: Yuri Shkuro <ys@...>
Cc: Matt Klein <mattklein123@...>, Alexis Richardson via cncf-toc <cncf-toc@...>, Gary Brown <gbrown@...>
Subject: Re: [cncf-toc] Jaeger graduation due diligence

 

Thanks Yuri and Matt.  This has been helpful for me.

 

I realised what was bothering me.  I think we should graduate Jaeger; but when we do so I think we should also provide some thoughts on the state of adoption.  Users will I hope feel more enlightened if we share that 1. The project is solid and has enough real use, 2. But the user base remains narrow, 3. Which will change as more ms apps are built and scaled, and tools like opentelemetry emerge, 4. And here is how you can help. 

 

 

 

On Mon, 7 Oct 2019, 21:44 Yuri Shkuro, <ys@...> wrote:

Thanks, Matt.

 

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

 

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

 

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

 

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:

Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

 

Weaveworks is not an enterprise, it is a small isv. 

 

Has jaeger found much traction among, say, non members of CNCF? 

 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  

 

 

 

 

On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

 

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.

 

 

On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:

Are you specifically referring to enterprise reselling/platform and not end-users?

 

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:

Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

 

Weaveworks are a happy user, as documented here.  

 

 

On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:

Hi folks,

 

Here is Jaeger's due diligence document for review prior to the graduation vote:

 

Please take a look at this document as well as the graduation PR and comment away!

 

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

 

Thanks,

Matt


Re: Jaeger graduation due diligence

William Morgan
 

On the topic of "the state of tracing", we JUST added tracing support to Linkerd. As someone with no real horse in the DT race, but with a desire to actually provide a concrete solution to our users, it was a particular painful exercise in navigating a complex landscape that at times seemed to consist solely of projects that were either deprecated or not ready for production. Writeup in https://linkerd.io/2019/10/07/a-guide-to-distributed-tracing-with-linkerd/.

(But Jaeger is great.)

-William


On Tue, Oct 8, 2019 at 8:03 AM via Lists.Cncf.Io <jbeda=vmware.com@...> wrote:

A couple of thoughts.

 

We should talk about the state of “tracing” in general.  It is useful to say that someone like Facebook is using it but it is also important to call out that they aren’t using Jaeger (right?).  Many many of the orgs doing this are very mature in their tooling and often have their own solutions in place.  Many more mature enterprises aren’t set up yet to really drive tracing.  Agree that OpenTelemetry could change that over time.

 

It is worth talking about the instrumentation/journey that many companies take.  Many start with metrics and logging (both structured and unstructured) long before they are mature enough for tracing.  I don’t kow if the CNCF has any advice for that.

 

Also worth noting that newer architectures that have more structure around user code (“serverless”) provide a great opportunity for automatically making tracing happen.  This will mature as open serverless frameworks become more mature.  (Things like Lambda are a closed system where they only integrate with their own tracing services).

 

Joe

 

From: <cncf-toc@...> on behalf of alexis richardson <alexis@...>
Date: Tuesday, October 8, 2019 at 5:45 AM
To: Yuri Shkuro <ys@...>
Cc: Matt Klein <mattklein123@...>, Alexis Richardson via cncf-toc <cncf-toc@...>, Gary Brown <gbrown@...>
Subject: Re: [cncf-toc] Jaeger graduation due diligence

 

Thanks Yuri and Matt.  This has been helpful for me.

 

I realised what was bothering me.  I think we should graduate Jaeger; but when we do so I think we should also provide some thoughts on the state of adoption.  Users will I hope feel more enlightened if we share that 1. The project is solid and has enough real use, 2. But the user base remains narrow, 3. Which will change as more ms apps are built and scaled, and tools like opentelemetry emerge, 4. And here is how you can help. 

 

 

 

On Mon, 7 Oct 2019, 21:44 Yuri Shkuro, <ys@...> wrote:

Thanks, Matt.

 

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

 

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

 

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

 

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:

Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

 

Weaveworks is not an enterprise, it is a small isv. 

 

Has jaeger found much traction among, say, non members of CNCF? 

 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  

 

 

 

 

On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

 

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.

 

 

On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:

Are you specifically referring to enterprise reselling/platform and not end-users?

 

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:

Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

 

Weaveworks are a happy user, as documented here.  

 

 

On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:

Hi folks,

 

Here is Jaeger's due diligence document for review prior to the graduation vote:

 

Please take a look at this document as well as the graduation PR and comment away!

 

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

 

Thanks,

Matt


Falco Incubation Proposal

Michael Ducy
 

CNCF TOC,

It's with great pleasure I present to you Falco's proposal to move to Incubation. Next Tuesday, October 15th, we will be presenting the project's annual review on the TOC call. In preparation for this call I've submitted Falco's formal Incubation proposal.

Please find the proposal here:


In summary:
 * All key metrics around contributions and committers have seen significant growth.
 * Downloads have seen significant growth.
 * The project has seen an increase in integrations, as well as Falco being incorporated into other products (Sumo Logic, Altran).
 * The community has seen increased participation and activity.
 * Public end users are documented in the projects' ADOPTERS.md. Two end users are speaking at Kubecon NA 2019.
 * The project completed a security audit and fixed 14 security bugs.
 * The project successfully participated in the Google Summer of Code.
 * The project successfully defined and shipped a roadmap.

We look forward to presenting on Tuesday!

Michael


Re: Jaeger graduation due diligence

Joe Beda <jbeda@...>
 

A couple of thoughts.

 

We should talk about the state of “tracing” in general.  It is useful to say that someone like Facebook is using it but it is also important to call out that they aren’t using Jaeger (right?).  Many many of the orgs doing this are very mature in their tooling and often have their own solutions in place.  Many more mature enterprises aren’t set up yet to really drive tracing.  Agree that OpenTelemetry could change that over time.

 

It is worth talking about the instrumentation/journey that many companies take.  Many start with metrics and logging (both structured and unstructured) long before they are mature enough for tracing.  I don’t kow if the CNCF has any advice for that.

 

Also worth noting that newer architectures that have more structure around user code (“serverless”) provide a great opportunity for automatically making tracing happen.  This will mature as open serverless frameworks become more mature.  (Things like Lambda are a closed system where they only integrate with their own tracing services).

 

Joe

 

From: <cncf-toc@...> on behalf of alexis richardson <alexis@...>
Date: Tuesday, October 8, 2019 at 5:45 AM
To: Yuri Shkuro <ys@...>
Cc: Matt Klein <mattklein123@...>, Alexis Richardson via cncf-toc <cncf-toc@...>, Gary Brown <gbrown@...>
Subject: Re: [cncf-toc] Jaeger graduation due diligence

 

Thanks Yuri and Matt.  This has been helpful for me.

 

I realised what was bothering me.  I think we should graduate Jaeger; but when we do so I think we should also provide some thoughts on the state of adoption.  Users will I hope feel more enlightened if we share that 1. The project is solid and has enough real use, 2. But the user base remains narrow, 3. Which will change as more ms apps are built and scaled, and tools like opentelemetry emerge, 4. And here is how you can help. 

 

 

 

On Mon, 7 Oct 2019, 21:44 Yuri Shkuro, <ys@...> wrote:

Thanks, Matt.

 

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

 

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

 

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

 

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:

Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

 

Weaveworks is not an enterprise, it is a small isv. 

 

Has jaeger found much traction among, say, non members of CNCF? 

 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  

 

 

 

 

On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

 

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:

I'm interested in end users who are also willing to self identify as enterprises.

 

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.

 

 

On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:

Are you specifically referring to enterprise reselling/platform and not end-users?

 

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:

Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

 

Weaveworks are a happy user, as documented here.  

 

 

On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:

Hi folks,

 

Here is Jaeger's due diligence document for review prior to the graduation vote:

 

Please take a look at this document as well as the graduation PR and comment away!

 

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

 

Thanks,

Matt


Re: Jaeger graduation due diligence

alexis richardson
 

Thanks Yuri and Matt.  This has been helpful for me.

I realised what was bothering me.  I think we should graduate Jaeger; but when we do so I think we should also provide some thoughts on the state of adoption.  Users will I hope feel more enlightened if we share that 1. The project is solid and has enough real use, 2. But the user base remains narrow, 3. Which will change as more ms apps are built and scaled, and tools like opentelemetry emerge, 4. And here is how you can help. 




On Mon, 7 Oct 2019, 21:44 Yuri Shkuro, <ys@...> wrote:
Thanks, Matt.

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:
My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:
Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

Weaveworks is not an enterprise, it is a small isv. 

Has jaeger found much traction among, say, non members of CNCF? 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  




On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

Yuri Shkuro <ys@...>
 

Thanks, Matt.

Yes, I wouldn't disagree about the niche qualification. Adopting tracing is relatively difficult today due to lack of reusable instrumentation, which will hopefully change with OpenTelemetry, and especially once the auto-instrumentation agents materialize. Having said that, many large tech companies use tracing, some of them highly effectively (e.g. Facebook). So I believe it's a niche that will only be getting bigger the more companies adopt microservices architectures.

On Mon, Oct 7, 2019 at 9:25 PM Matt Klein <mattklein123@...> wrote:
My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:
Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

Weaveworks is not an enterprise, it is a small isv. 

Has jaeger found much traction among, say, non members of CNCF? 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  




On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

Matt Klein
 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.

I will let @Yuri Shkuro and @Gary Brown chime in, but FWIW, my own take is that distributed tracing in general is still fairly niche and likely to be found at early adopter companies/enterprises. I'm not sure that we can expect "wide adoption" in this space. However, in my personal opinion, Jaeger's adoption is substantial enough to merit graduation.

On Mon, Oct 7, 2019 at 9:01 PM Alexis Richardson <alexis@...> wrote:
Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

Weaveworks is not an enterprise, it is a small isv. 

Has jaeger found much traction among, say, non members of CNCF? 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  




On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

alexis richardson
 

Ticketmaster is an enterprise.  So is Uber but they are the parent Co. 

Weaveworks is not an enterprise, it is a small isv. 

Has jaeger found much traction among, say, non members of CNCF? 

My worry is that somehow the product is quite niche.  It would be good to have broad adoption.  




On Mon, 7 Oct 2019, 20:55 Matt Klein, <mattklein123@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

Matt Klein
 

I'm interested in end users who are also willing to self identify as enterprises.

We are probably about to get way off topic, but just to make sure I understand, what exactly is the definition of an enterprise here? Is Ticketmaster an enterprise? Is Uber?

On Mon, Oct 7, 2019 at 8:43 PM Alexis Richardson <alexis@...> wrote:
I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

alexis richardson
 

I'm interested in end users who are also willing to self identify as enterprises.

Prometheus has a lot of these.  I had hoped Jaeger would be similar. 

Again it's not a deal breaker for me at all, but I'd love to hear there are enterprise users.


On Mon, 7 Oct 2019, 20:36 Matt Klein, <mattklein123@...> wrote:
Are you specifically referring to enterprise reselling/platform and not end-users?

On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

Matt Klein
 

Are you specifically referring to enterprise reselling/platform and not end-users?


On Mon, Oct 7, 2019 at 8:27 PM Alexis Richardson <alexis@...> wrote:
Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Jaeger graduation due diligence

alexis richardson
 

Not that this should be a blocker but how much evidence of enterprise or CSP use is Jaeger seeing?

Weaveworks are a happy user, as documented here.  


On Mon, 7 Oct 2019, 20:23 Matt Klein, <mattklein123@...> wrote:
Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Jaeger graduation due diligence

Matt Klein
 

Hi folks,

Here is Jaeger's due diligence document for review prior to the graduation vote:

Please take a look at this document as well as the graduation PR and comment away!

Thanks to Yuri and Gary for their hard work helping to put all of the material together.

Thanks,
Matt


Re: Next steps for KUDO

Tobias Knaup
 

OK thanks, we opened the issue here and would like to present this Wednesday if possible: https://github.com/cncf/sig-app-delivery/issues/8.
Is this new process documented somewhere? It's hard to figure out which steps to follow and we keep getting different answers depending on who we're talking to. For example I see other project presentations on the agenda for Wednesday but no corresponding issues on Github. How did those get on the agenda?
Also in our case Xiang and Brian from the TOC already did a review so can someone clarify the scope of this additional review?

Thanks,

Tobi


On Mon, Oct 7, 2019 at 13:20 Amye Scavarda Perrin <amye@...> wrote:
So we are working on getting the SIGs more engaged in the process of
helping to review projects.
https://github.com/cncf/sig-app-delivery/issues where we should be
working to get SIG App-Delivery to review, go ahead and open an issue
there to get on their agenda.

On Mon, Oct 7, 2019 at 1:09 PM Tobias Knaup <tobi@...> wrote:
>
> All,
>
> The KUDO project presented at the TOC meeting in mid September and we were wondering what the next steps are. At the end of our presentation there was talk that SIG App Delivery would get involved but we haven’t heard anything since. Can someone clarify the process and if there’s a next step the project should take?
>
> Thanks,
>
> Tobi


--
Amye Scavarda Perrin | Program Manager, CNCF | amye@...


Re: Next steps for KUDO

Amye Scavarda Perrin
 

So we are working on getting the SIGs more engaged in the process of
helping to review projects.
https://github.com/cncf/sig-app-delivery/issues where we should be
working to get SIG App-Delivery to review, go ahead and open an issue
there to get on their agenda.

On Mon, Oct 7, 2019 at 1:09 PM Tobias Knaup <tobi@...> wrote:

All,

The KUDO project presented at the TOC meeting in mid September and we were wondering what the next steps are. At the end of our presentation there was talk that SIG App Delivery would get involved but we haven’t heard anything since. Can someone clarify the process and if there’s a next step the project should take?

Thanks,

Tobi
--
Amye Scavarda Perrin | Program Manager, CNCF | amye@...


Next steps for KUDO

Tobias Knaup
 

All,

The KUDO project presented at the TOC meeting in mid September and we were wondering what the next steps are. At the end of our presentation there was talk that SIG App Delivery would get involved but we haven’t heard anything since. Can someone clarify the process and if there’s a next step the project should take?

Thanks,

Tobi


TUF graduation due diligence document...

Liz Rice
 

Sending round the TOC for any initial thoughts
---------- Forwarded message ----------
From: Justin Cappos <justincappos@...>
Date: 5 Oct 2019, 13:26 +0100
To: Liz Rice <liz@...>
Cc: Amye Scavarda Perrin <amye@...>, Justin Cormack <justin.cormack@...>, Santiago Torres <santiago@...>, Lukas Pühringer <lukas.puehringer@...>, Trishank Kuppusamy <trishank.kuppusamy@...>
Subject: TUF graduation due diligence document...

Amye had mentioned to us that the due diligence document is holding up a TUF graduation vote.  I've included a first draft of the TUF graduation due diligence document...


I'd be happy to iterate with you or others on the document or to provide anything else needed to bring TUF up for a vote in the TOC.

Thanks,
Justin


Draft Charter for CNCF Network SIG

Lee Calcote
 

Hi,

Since inception of CNCF SIGs, we’ve intended to breathe new life into the Networking WG, reincarnating it as a SIG with a redefined charter. Its draft charter is now ready for broad review and community feedback. If you have an interest in this area, please reach out to me, Matt or Ken and/or comment in the document.

The formation of this SIG will be discussed briefly in today’s TOC call. Our intention is have this SIG formed in the next couple weeks. Those who’s signaled interest have been subscribed to the mailing list below. For any others interested, join at-will:

Slack Channel (#sig-network)
Join SIG-Network mailer at lists.cncf.io

Regards,
- Lee



Re: TOC Meeting 10/1/2019

Brendan Burns
 

Likewise, I need to miss, apologies!



From: cncf-toc@... <cncf-toc@...> on behalf of Matt Klein via Lists.Cncf.Io <mattklein123=gmail.com@...>
Sent: Tuesday, October 1, 2019 7:28 AM
To: Amye Scavarda Perrin <ascavarda@...>
Cc: cncf-toc@... <cncf-toc@...>
Subject: Re: [cncf-toc] TOC Meeting 10/1/2019
 
Apologies but I will miss the meeting today due to travel.

On Tue, Oct 1, 2019 at 7:17 AM Amye Scavarda Perrin <ascavarda@...> wrote:

Hi all! 
We'll be meeting today: 
https://docs.google.com/document/d/1jpoKT12jf2jTf-2EJSAl4iTdA7Aoj_uiI19qIaECNFc/edit#

We'll be getting updates from the SIGs, a follow-up presentation from Longhorn, Jaeger's Graduation Review and a SIG-Security Review: Open Policy Agent. 
Moving forward between now and November 5th, we'll be prioritizing graduation reviews. 
--
Amye Scavarda Perrin | Program Manager, CNCF | amye@...