Prometheus metrics

85 views
Skip to first unread message

Daniel Slavetskiy

unread,
Sep 28, 2020, 11:02:34 AM9/28/20
to UniMRCP
Hi Arsen,

do you have any plans in your roadmap to implement Promeheus (https://prometheus.io/) support?

Arsen Chaloyan

unread,
Oct 13, 2020, 10:49:06 PM10/13/20
to UniMRCP
Hi Daniel,

We set up a test bed to explore the Prometheus toolkit a month or two ago and quite liked the capabilities of this framework. Implementation of Prometheus metrics instrumentation has been in progress for one of private plugins and will be complete soon.

I cannot tell whether or not the same is going to be implemented for all the public plugins. It would be nice to have a single Prometheus exposure implemented commonly for UniMRCP server as opposed to individual plugins. The generic approach has a larger scope effort-wise, though.


On Mon, Sep 28, 2020 at 8:02 AM Daniel Slavetskiy <daniel.s...@lekta.ai> wrote:
Hi Arsen,

do you have any plans in your roadmap to implement Promeheus (https://prometheus.io/) support?

--
You received this message because you are subscribed to the Google Groups "UniMRCP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unimrcp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unimrcp/e363cefd-ee84-4507-932b-ddbfec0a093fn%40googlegroups.com.


--
Arsen Chaloyan
Author of UniMRCP
http://www.unimrcp.org

Daniel Slavetskiy

unread,
Oct 29, 2020, 7:17:49 AM10/29/20
to UniMRCP
Hi Arsen, 

thanks for reply, yes the generic approach is better. But if you don't want to stick to specific impl and data format as there are different monitoring tools, as a first step it would be great to expose different stats through callbacks in plugin interface, then it's quite universal and gives the freedom and responsibility to impl the integration with whatever tool.

I've also made a separate thread(https://groups.google.com/g/unimrcp/c/PzOId63txuI/m/PWMFRmUvBwAJ) for RTCP stats, but it's connected with this one

WDYT?

Arsen Chaloyan

unread,
Nov 3, 2020, 7:53:22 PM11/3/20
to UniMRCP
Hi Daniel,

By referring to the generic approach, I meant to have a status/statistics interface in the core which could be implemented for Prometheus, Icinga or any other tool in an external module. The core should definitely be independent of final implementation.

There is nothing wrong having the Prometheus exposure implemented in the plugin on one hand. On the other, what if you have multiple plugins loaded into a single UniMRCP server instance? All the plugins would have to link with the same library and duplicate the code for implementation of exposure.

On Thu, Oct 29, 2020 at 4:17 AM Daniel Slavetskiy <daniel.s...@lekta.ai> wrote:
Hi Arsen, 

thanks for reply, yes the generic approach is better. But if you don't want to stick to specific impl and data format as there are different monitoring tools, as a first step it would be great to expose different stats through callbacks in plugin interface, then it's quite universal and gives the freedom and responsibility to impl the integration with whatever tool.

I've also made a separate thread(https://groups.google.com/g/unimrcp/c/PzOId63txuI/m/PWMFRmUvBwAJ) for RTCP stats, but it's connected with this one

WDYT?

On Wednesday, 14 October 2020 at 04:49:06 UTC+2 Arsen Chaloyan wrote:
Hi Daniel,

We set up a test bed to explore the Prometheus toolkit a month or two ago and quite liked the capabilities of this framework. Implementation of Prometheus metrics instrumentation has been in progress for one of private plugins and will be complete soon.

I cannot tell whether or not the same is going to be implemented for all the public plugins. It would be nice to have a single Prometheus exposure implemented commonly for UniMRCP server as opposed to individual plugins. The generic approach has a larger scope effort-wise, though.


On Mon, Sep 28, 2020 at 8:02 AM Daniel Slavetskiy <daniel.s...@lekta.ai> wrote:
Hi Arsen,

do you have any plans in your roadmap to implement Promeheus (https://prometheus.io/) support?

--
You received this message because you are subscribed to the Google Groups "UniMRCP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unimrcp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unimrcp/e363cefd-ee84-4507-932b-ddbfec0a093fn%40googlegroups.com.


--
Arsen Chaloyan
Author of UniMRCP
http://www.unimrcp.org

--
You received this message because you are subscribed to the Google Groups "UniMRCP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unimrcp+u...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages