any in-progress Redfish TelemetryService enhancements?

any in-progress Redfish TelemetryService enhancements?

[Brad Bishop]

Has anyone thought about how to implement the Append behaviors for the 
ReportUpdates property in a MetricReportDefinition?

thx - brad

Re: any in-progress Redfish TelemetryService enhancements?

[Ed Tanous]

On Thu, Feb 18, 2021 at 11:16 AM Brad Bishop
<bradleyb@fuzziesquirrel.com> wrote:
>
> Has anyone thought about how to implement the Append behaviors for the
> ReportUpdates property in a MetricReportDefinition?
>

As an aside, if anyone else is interested in Telemetry Service,
there's a lot of code there that's not having a lot of community
feedback or testing on the reviews, so if anyone else is interested
please pull down the patches and provide constructive feedback.

https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+(status:open+OR+status:merged)+status:open

> thx - brad

Re: any in-progress Redfish TelemetryService enhancements?

[Brad Bishop]

On Thu, Feb 18, 2021 at 12:04:57PM -0800, Ed Tanous wrote:
>On Thu, Feb 18, 2021 at 11:16 AM Brad Bishop
><bradleyb@fuzziesquirrel.com> wrote:
>>
>> Has anyone thought about how to implement the Append behaviors for the
>> ReportUpdates property in a MetricReportDefinition?
>>
>
>As an aside, if anyone else is interested in Telemetry Service,
>there's a lot of code there that's not having a lot of community
>feedback or testing on the reviews, so if anyone else is interested
>please pull down the patches and provide constructive feedback.
>
>https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+(status:open+OR+status:merged)+status:open

I pulled these in and did some quick sanity checking.  I was able to 
create a report definition and then see the report get generated and 
read it.

I have to say - it was really nice being able to pull a feature in and 
having it work out of the box.  So really great job to the Intel team in 
that respect.

I did find a bug with the telemetry service itself - when reports with 
multiple metrics are created, the last (or maybe all but the first?) 
metric in the report isn't updated after the initial reading.  I was too 
lazy to file a bug and didn't save any information.  I'll try to do that 
soon.

Re: any in-progress Redfish TelemetryService enhancements?

[Wludzik, Jozef]

Hi, "Append" is on the list of to dos and should be ready till summer
(might be ready). Other feature that are described by Redfish Telemetry
Model should be ready till the end of summer too -> triggers and
collection functions.

About the issue, the metric reading is not updated when there is no
update from the sensor dbus. To debug it best way to monitor the
PropertiesChanged signal from the service that implement the sensor. If
there is no signal from service than metric won't be updated. If there
is a signal, then Telemetry service has a regression.

I would be very glad for any feedback on telemetry, even filing a bug on
Telemetry sounds great. Thanks in advance

Thanks,
Jozef

On 2/19/2021 20:33 PM, Brad Bishop wrote:
> On Thu, Feb 18, 2021 at 12:04:57PM -0800, Ed Tanous wrote:
>> On Thu, Feb 18, 2021 at 11:16 AM Brad Bishop
>> <bradleyb@fuzziesquirrel.com> wrote:
>>>
>>> Has anyone thought about how to implement the Append behaviors for the
>>> ReportUpdates property in a MetricReportDefinition?
>>>
>>
>> As an aside, if anyone else is interested in Telemetry Service,
>> there's a lot of code there that's not having a lot of community
>> feedback or testing on the reviews, so if anyone else is interested
>> please pull down the patches and provide constructive feedback.
>>
>> https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+(status:open+OR+status:merged)+status:open
>>
> 
> I pulled these in and did some quick sanity checking.  I was able to
> create a report definition and then see the report get generated and
> read it.
> 
> I have to say - it was really nice being able to pull a feature in and
> having it work out of the box.  So really great job to the Intel team in
> that respect.
> 
> I did find a bug with the telemetry service itself - when reports with
> multiple metrics are created, the last (or maybe all but the first?)
> metric in the report isn't updated after the initial reading.  I was too
> lazy to file a bug and didn't save any information.  I'll try to do that
> soon.

Re: any in-progress Redfish TelemetryService enhancements?

[Brad Bishop]

On Wed, Feb 24, 2021 at 01:09:56PM +0100, Wludzik, Jozef wrote:
>Hi, "Append" is on the list of to dos and should be ready till summer
>(might be ready). 

That's great!  Any chance you could you provide any hints on how you 
expect it to be implemented - I'm not sure if I can wait that long to 
get started.  I would like to make sure that if I start working on it, 
it will have your approval from a high level view.  If you have not 
given it any thought, no problem - I'll come up with a proposal and 
report back here.

Re: any in-progress Redfish TelemetryService enhancements?

[George Liu]

Hi, Wludzik, Brad:

Since we have some open-issues that need to rely on Telemetry service:
https://github.com/ibm-openbmc/dev/issues/2968
https://github.com/ibm-openbmc/dev/issues/2969

I want to make sure that Intel team will finish it by summer?
If it is, that is great, we will always focus on and participate in
the code-review.
Otherwise, as Brad said, we will come up with a proposal and report back here.

Thanks,
George Liu

On Mon, Mar 1, 2021 at 11:06 PM Brad Bishop <bradleyb@fuzziesquirrel.com> wrote:
>
> On Wed, Feb 24, 2021 at 01:09:56PM +0100, Wludzik, Jozef wrote:
> >Hi, "Append" is on the list of to dos and should be ready till summer
> >(might be ready).
>
> That's great!  Any chance you could you provide any hints on how you
> expect it to be implemented - I'm not sure if I can wait that long to
> get started.  I would like to make sure that if I start working on it,
> it will have your approval from a high level view.  If you have not
> given it any thought, no problem - I'll come up with a proposal and
> report back here.

Re: any in-progress Redfish TelemetryService enhancements?

[Ed Tanous]

On Tue, Mar 9, 2021 at 7:23 PM George Liu <liuxiwei1013@gmail.com> wrote:
>
> Hi, Wludzik, Brad:
>
> Since we have some open-issues that need to rely on Telemetry service:
> https://github.com/ibm-openbmc/dev/issues/2968
> https://github.com/ibm-openbmc/dev/issues/2969

The people that care about those things should really be reviewing the
patches that are already in flight.  Unless I'm mistaken, I see no
reviews from George on any of the telemetry patches:
https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+reviewedby:liuxiwei%2540inspur.com

>
> I want to make sure that Intel team will finish it by summer?
> If it is, that is great, we will always focus on and participate in
> the code-review.
> Otherwise, as Brad said, we will come up with a proposal and report back here.

You should participate in the reviews and help test either way.  If
the initial feature doesn't land on master, there's no point in
working on or planning secondary features.

>
> Thanks,
> George Liu
>
> On Mon, Mar 1, 2021 at 11:06 PM Brad Bishop <bradleyb@fuzziesquirrel.com> wrote:
> >
> > On Wed, Feb 24, 2021 at 01:09:56PM +0100, Wludzik, Jozef wrote:
> > >Hi, "Append" is on the list of to dos and should be ready till summer
> > >(might be ready).
> >
> > That's great!  Any chance you could you provide any hints on how you
> > expect it to be implemented - I'm not sure if I can wait that long to
> > get started.  I would like to make sure that if I start working on it,
> > it will have your approval from a high level view.  If you have not
> > given it any thought, no problem - I'll come up with a proposal and
> > report back here.

Re: any in-progress Redfish TelemetryService enhancements?

[Brad Bishop]

On Tue, Mar 09, 2021 at 09:07:51PM -0800, Ed Tanous wrote:
>On Tue, Mar 9, 2021 at 7:23 PM George Liu <liuxiwei1013@gmail.com> wrote:
>>
>> Hi, Wludzik, Brad:
>>
>> Since we have some open-issues that need to rely on Telemetry service:
>> https://github.com/ibm-openbmc/dev/issues/2968
>> https://github.com/ibm-openbmc/dev/issues/2969
>
>The people that care about those things should really be reviewing the
>patches that are already in flight.  Unless I'm mistaken, I see no
>reviews from George on any of the telemetry patches:
>https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+reviewedby:liuxiwei%2540inspur.com

No disagreement that those who care about telemetry should help review 
telemetry patches.  

For what its worth I asked George to post to this thread so that 
everyone is aware of his intent to work on these features.  OpenBMC is a 
siloed project and I'm trying to set an example and break the silos down 
by overcommunicating and asking others to do the same.

>> I want to make sure that Intel team will finish it by summer?
>> If it is, that is great, we will always focus on and participate in
>> the code-review.
>> Otherwise, as Brad said, we will come up with a proposal and report back here.
>
>You should participate in the reviews and help test either way.

I think there is a misunderstanding here.  The point of this thread was 
simply to find out if the telemetry experts (or anyone else that cares 
to comment) had any opinions on how these functions should be 
implemented.  Clearly George can't review his own proposal can he?

>If the initial feature doesn't land on master, there's no point in
>working on or planning secondary features.

I understand you are focused on reviews and for good reason.  Does it 
not make sense to work on reviews and design/planning in parallel?

-brad

Re: any in-progress Redfish TelemetryService enhancements?

[Ed Tanous]

On Wed, Mar 10, 2021 at 5:52 AM Brad Bishop <bradleyb@fuzziesquirrel.com> wrote:
>
> On Tue, Mar 09, 2021 at 09:07:51PM -0800, Ed Tanous wrote:
> >On Tue, Mar 9, 2021 at 7:23 PM George Liu <liuxiwei1013@gmail.com> wrote:
> >>
> >> Hi, Wludzik, Brad:
> >>
> >> Since we have some open-issues that need to rely on Telemetry service:
> >> https://github.com/ibm-openbmc/dev/issues/2968
> >> https://github.com/ibm-openbmc/dev/issues/2969
> >
> >The people that care about those things should really be reviewing the
> >patches that are already in flight.  Unless I'm mistaken, I see no
> >reviews from George on any of the telemetry patches:
> >https://gerrit.openbmc-project.xyz/q/topic:%2522telemetry%2522+reviewedby:liuxiwei%2540inspur.com
>
> No disagreement that those who care about telemetry should help review
> telemetry patches.
>
> For what its worth I asked George to post to this thread so that
> everyone is aware of his intent to work on these features.  OpenBMC is a
> siloed project and I'm trying to set an example and break the silos down
> by overcommunicating and asking others to do the same.

I have no issue with posting here;  That's great, and helps break down
silos, where I struggle is where we propose features on top of the
existing.

>
> >> I want to make sure that Intel team will finish it by summer?
> >> If it is, that is great, we will always focus on and participate in
> >> the code-review.
> >> Otherwise, as Brad said, we will come up with a proposal and report back here.
> >
> >You should participate in the reviews and help test either way.
>
> I think there is a misunderstanding here.  The point of this thread was
> simply to find out if the telemetry experts (or anyone else that cares
> to comment) had any opinions on how these functions should be
> implemented.  Clearly George can't review his own proposal can he?

I don't think any would argue that, but if George needs someone to
review his proposal, he should be reviewing other proposals, right?

>
> >If the initial feature doesn't land on master, there's no point in
> >working on or planning secondary features.
>
> I understand you are focused on reviews and for good reason.  Does it
> not make sense to work on reviews and design/planning in parallel?
>

There's no problem with working on things in parallel, but when many
patches have been in review for a very long time, with almost no input
outside of maintainers and the submitters have asked for help
reviewing multiple times, it's hard when someone pops up months later
and suggests we stack more patches on top, while at the same time
ignoring others patches.

George, to be clear, this is not to single you out, and this problem
goes beyond you.  To some extent, the email just hit a bit of a nerve.
It would be really helpful if you could add your input and participate
in the reviews.  In this specific case, I recall pointing out the
missing Min/Max/Average aggregations (from the first bug) in one of
the original telemetry code reviews.  I don't remember what the answer
was there, but I don't think it got added.  I don't think append
behavior has come up, but I suspect it would've changed some of the
initial design if we had known it was desired.

> -brad

Re: any in-progress Redfish TelemetryService enhancements?

[Brad Bishop]

On Wed, Mar 10, 2021 at 07:12:39AM -0800, Ed Tanous wrote:

>It would be really helpful if you could add your input and participate
>in the reviews.  In this specific case, I recall pointing out the
>missing Min/Max/Average aggregations (from the first bug) in one of
>the original telemetry code reviews.  I don't remember what the answer
>was there, but I don't think it got added.  I don't think append
>behavior has come up, but I suspect it would've changed some of the
>initial design if we had known it was desired.

FWIW this interest in the telemetry work has only emerged very recently 
(around when I sent the first mail in this thread).

-brad

Re: any in-progress Redfish TelemetryService enhancements?

[George Liu]

Hi All:
Around Telemetry(Support AppendWrapsWhenFull):
I added a new interface and merged:
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-dbus-interfaces/+/41010

So, @Wludzik, Wonder if there is a plan to implement it?
If Yes, Could you share the time cut-off point?

Thanks again.

George Liu

On Wed, Mar 10, 2021 at 11:47 PM Brad Bishop
<bradleyb@fuzziesquirrel.com> wrote:
>
> On Wed, Mar 10, 2021 at 07:12:39AM -0800, Ed Tanous wrote:
>
> >It would be really helpful if you could add your input and participate
> >in the reviews.  In this specific case, I recall pointing out the
> >missing Min/Max/Average aggregations (from the first bug) in one of
> >the original telemetry code reviews.  I don't remember what the answer
> >was there, but I don't think it got added.  I don't think append
> >behavior has come up, but I suspect it would've changed some of the
> >initial design if we had known it was desired.
>
> FWIW this interest in the telemetry work has only emerged very recently
> (around when I sent the first mail in this thread).
>
> -brad

Re: any in-progress Redfish TelemetryService enhancements?

[Grobelny, Krzysztof]

[-- Attachment #1: Type: text/plain, Size: 2357 bytes --]

Hello,

Our current plan looks like this:

Completing outgoing review in bmcweb for telemetry:
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/33363 
<https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/33363>
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/38798/27 
<https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/38798/27>
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/41813/1 
<https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/41813/1>

With start of July will finish work on features already started and fix 
know issues, then we will start working on "Append".
By currently ongoing feature I mean Triggers and Aggregation functions, 
review below doesn't cover everything we need to implement:
https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/41064 
<https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/41064>, 
https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/40941 
<https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/40941>, 
https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/41678 
<https://gerrit.openbmc-project.xyz/c/openbmc/telemetry/+/41678>

My estimate is that we will deliver "Append" in Q4 2021.

Best regards,
Krzysztof Grobelny

On 2021-04-28 04:15, George Liu wrote:
> Hi All:
> Around Telemetry(Support AppendWrapsWhenFull):
> I added a new interface and merged:
> https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-dbus-interfaces/+/41010
>
> So, @Wludzik, Wonder if there is a plan to implement it?
> If Yes, Could you share the time cut-off point?
>
> Thanks again.
>
> George Liu
>
> On Wed, Mar 10, 2021 at 11:47 PM Brad Bishop
> <bradleyb@fuzziesquirrel.com> wrote:
>> On Wed, Mar 10, 2021 at 07:12:39AM -0800, Ed Tanous wrote:
>>
>>> It would be really helpful if you could add your input and participate
>>> in the reviews.  In this specific case, I recall pointing out the
>>> missing Min/Max/Average aggregations (from the first bug) in one of
>>> the original telemetry code reviews.  I don't remember what the answer
>>> was there, but I don't think it got added.  I don't think append
>>> behavior has come up, but I suspect it would've changed some of the
>>> initial design if we had known it was desired.
>> FWIW this interest in the telemetry work has only emerged very recently
>> (around when I sent the first mail in this thread).
>>
>> -brad

[-- Attachment #2: Type: text/html, Size: 5249 bytes --]