RE: Sling Health Checks

[Jason Bailey]

I appreciate the feedback. There is so much churn in the developer mailing list 
I found myself unable to track new additions or changes like I used to do.

The touch point with is AppDynamics is JMX, which I see is supported in the 
dropwizard metrics.

For the issue I had, I am fortunate to have a talented intern and she was able 
to get everything working for our last Sprint. Going forward, I will be 
floating this functionality around and see if we can't start leveraging it soon.

Thanks
-Jason

-Original Message-
From: Björn Weide [mailto:bjoern.we...@netcentric.biz] 
Sent: Tuesday, February 07, 2017 4:12 AM
To: users@sling.apache.org
Cc: georg.henz...@netcentric.biz
Subject: Re: Sling Health Checks

Just my 2cents on that one - for me it sounds like it would be a good fit to 
use Sling Metrics here to record your metrics (like # of submissions in your 
case). You could then use a custom reporter to expose the data to AppDynamics 
(not sure about the integration points there).
In case you want to use a HC for that metrics you could implement some kind of 
delta reporter that simply reports the deltas of selected metrics to an 
internal store and then implement a HC checking these deltas against your 
thresholds (we are using this setup in one of our projects).

Best regards,
Bjoern

On Tue, Jan 31, 2017 at 4:59 PM, Jason Bailey <jason.bai...@sas.com> wrote:

> So I'm part of a team that I will characterize as being "highly 
> resource optimized" We have an AppDynamics instance that we use to 
> monitor the state of our instances as much as possible. One of the 
> OOTB features is the ability to tie into JMX and present graphs and 
> historical information on specific items.
>
> We have a variety of business processes and functions in Sling that 
> would be beneficial to us to monitor, and in particular graph, so that 
> we can do intelligent alarming.  Where we are having problems is that 
> collection of state that would allow us to do intelligent monitoring.
>
> Example being submission processing. If the number of submissions each 
> instance is handling falls below a certain threshold or increases 
> beyond a certain number, it's an issue of concern, a warning. But 
> certain days and a particular time, it's a situation where need 
> someone looking at it immediately.
>
> Additionally having a value being returned would also allow us to 
> track trends or unusual events in a historical manner so that we can 
> hopefully identify issues before they happen.
>
> -Jason
>
> -Original Message-
> From: Georg Henzler [mailto:slin...@ghenzler.de]
> Sent: Tuesday, January 31, 2017 3:24 AM
> To: users@sling.apache.org
> Subject: Re: Sling Health Checks
>
> Hi Jason,
>
> no it is not possible at the moment - but nobody stops you to make the 
> HC component also provide an additional, custom MBean that shares some 
> additional result properties of the HC execution to other services.
>
> In theory the HC result could be extended to allow for custom 
> properties (that are then in turn exposed via JMX), but the idea 
> behind the HCs is that due to the clearly defined result type (with a 
> clear semantic for each
> status) consumers can easily decide whether an instance is healthy or 
> not, additional properties might be confusing. Could you describe your 
> use case in a bit more detail?
>
> Regards
> Georg
>
> On 2017-01-30 18:56, Jason Bailey wrote:
> > Anyone know if it's possible to use the Sling Health Checks to 
> > expose a value via the MBean other than the ones explicitly defined 
> > in the Result Object?
> >
> > Thanks
> > -Jason
>

Re: Sling Health Checks

[Björn Weide]

Just my 2cents on that one - for me it sounds like it would be a good fit
to use Sling Metrics here to record your metrics (like # of submissions in
your case). You could then use a custom reporter to expose the data to
AppDynamics (not sure about the integration points there).
In case you want to use a HC for that metrics you could implement some kind
of delta reporter that simply reports the deltas of selected metrics to an
internal store and then implement a HC checking these deltas against your
thresholds (we are using this setup in one of our projects).

Best regards,
Bjoern

On Tue, Jan 31, 2017 at 4:59 PM, Jason Bailey <jason.bai...@sas.com> wrote:

> So I'm part of a team that I will characterize as being "highly resource
> optimized" We have an AppDynamics instance that we use to monitor the state
> of our instances as much as possible. One of the OOTB features is the
> ability to tie into JMX and present graphs and historical information on
> specific items.
>
> We have a variety of business processes and functions in Sling that would
> be beneficial to us to monitor, and in particular graph, so that we can do
> intelligent alarming.  Where we are having problems is that collection of
> state that would allow us to do intelligent monitoring.
>
> Example being submission processing. If the number of submissions each
> instance is handling falls below a certain threshold or increases beyond a
> certain number, it's an issue of concern, a warning. But certain days and a
> particular time, it's a situation where need someone looking at it
> immediately.
>
> Additionally having a value being returned would also allow us to track
> trends or unusual events in a historical manner so that we can hopefully
> identify issues before they happen.
>
> -Jason
>
> -Original Message-
> From: Georg Henzler [mailto:slin...@ghenzler.de]
> Sent: Tuesday, January 31, 2017 3:24 AM
> To: users@sling.apache.org
> Subject: Re: Sling Health Checks
>
> Hi Jason,
>
> no it is not possible at the moment - but nobody stops you to make the HC
> component also provide an additional, custom MBean that shares some
> additional result properties of the HC execution to other services.
>
> In theory the HC result could be extended to allow for custom properties
> (that are then in turn exposed via JMX), but the idea behind the HCs is
> that due to the clearly defined result type (with a clear semantic for each
> status) consumers can easily decide whether an instance is healthy or not,
> additional properties might be confusing. Could you describe your use case
> in a bit more detail?
>
> Regards
> Georg
>
> On 2017-01-30 18:56, Jason Bailey wrote:
> > Anyone know if it's possible to use the Sling Health Checks to expose
> > a value via the MBean other than the ones explicitly defined in the
> > Result Object?
> >
> > Thanks
> > -Jason
>

RE: Sling Health Checks

[Jason Bailey]

Yes, you are right. 

I have quite a few scenarios where beyond the indicator that something is 
wrong, that the healthcheck provides me, I need to capture a value that to 
represent the scale of the problem. So that I can have analysis. I was hoping 
to have the option to provide that as part of the healthcheck rather than 
writing  two separate things. :)

All good.

-Original Message-
From: Bertrand Delacretaz [mailto:bdelacre...@apache.org] 
Sent: Tuesday, January 31, 2017 11:59 AM
To: users <users@sling.apache.org>
Subject: Re: Sling Health Checks

Hi,

On Tue, Jan 31, 2017 at 4:59 PM, Jason Bailey <jason.bai...@sas.com> wrote:
> ...We have an AppDynamics instance that we use to monitor the state of our 
> instances as much as possible...

> ...If the number of submissions each instance is handling falls below 
> a certain threshold or increases beyond a certain number, it's an 
> issue of concern, a warning. But certain days and a particular time, it's a 
> situation where need someone looking at it immediately...

Isn't this a case for an MBean that just reports the number of submissions, and 
a health check that checks that they are within range, maybe based on days and 
times? I think that would work with our standard health checks modules.

-Bertrand

RE: Sling Health Checks

[Jason Bailey]

So I'm part of a team that I will characterize as being "highly resource 
optimized" We have an AppDynamics instance that we use to monitor the state of 
our instances as much as possible. One of the OOTB features is the ability to 
tie into JMX and present graphs and historical information on specific items.

We have a variety of business processes and functions in Sling that would be 
beneficial to us to monitor, and in particular graph, so that we can do 
intelligent alarming.  Where we are having problems is that collection of state 
that would allow us to do intelligent monitoring. 

Example being submission processing. If the number of submissions each instance 
is handling falls below a certain threshold or increases beyond a certain 
number, it's an issue of concern, a warning. But certain days and a particular 
time, it's a situation where need someone looking at it immediately.

Additionally having a value being returned would also allow us to track trends 
or unusual events in a historical manner so that we can hopefully identify 
issues before they happen.

-Jason

-Original Message-
From: Georg Henzler [mailto:slin...@ghenzler.de] 
Sent: Tuesday, January 31, 2017 3:24 AM
To: users@sling.apache.org
Subject: Re: Sling Health Checks

Hi Jason,

no it is not possible at the moment - but nobody stops you to make the HC 
component also provide an additional, custom MBean that shares some additional 
result properties of the HC execution to other services.

In theory the HC result could be extended to allow for custom properties (that 
are then in turn exposed via JMX), but the idea behind the HCs is that due to 
the clearly defined result type (with a clear semantic for each status) 
consumers can easily decide whether an instance is healthy or not, additional 
properties might be confusing. Could you describe your use case in a bit more 
detail?

Regards
Georg

On 2017-01-30 18:56, Jason Bailey wrote:
> Anyone know if it's possible to use the Sling Health Checks to expose 
> a value via the MBean other than the ones explicitly defined in the 
> Result Object?
> 
> Thanks
> -Jason

Re: Sling Health Checks

[Bertrand Delacretaz]

Hi.

On Tue, Jan 31, 2017 at 9:23 AM, Georg Henzler  wrote:
> ...nobody stops you to make the HC
> component also provide an additional, custom MBean that shares some
> additional result properties of the HC execution to other services

Or maybe put that additional information in its own MBean, and use HC
just to check that those values are in range - that might be a better
separation of concerns.

-Bertrand

Re: Sling Health Checks

[Georg Henzler]

Hi Jason,

no it is not possible at the moment - but nobody stops you to make the 
HC component also provide an additional, custom MBean that shares some 
additional result properties of the HC execution to other services.


In theory the HC result could be extended to allow for custom properties 
(that are then in turn exposed via JMX), but the idea behind the HCs is 
that due to the clearly defined result type (with a clear semantic for 
each status) consumers can easily decide whether an instance is healthy 
or not, additional properties might be confusing. Could you describe 
your use case in a bit more detail?


Regards
Georg

On 2017-01-30 18:56, Jason Bailey wrote:

Anyone know if it's possible to use the Sling Health Checks to expose
a value via the MBean other than the ones explicitly defined in the
Result Object?

Thanks
-Jason