Send VM Metrics directly to Grafana

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Send VM Metrics directly to Grafana

Frank Muller
Hi guys

While deploying one of our Erlang products in a highly secure environment (private bank), this partner required from us to export the VM metrics to Grafana. 

I know that a Prometheus Erlang client do exist, but our partner refused to deploy Prometheus (didn’t pass their pentesting phase).

Question: is there a way to send the VM metrics directly to Grafana?
Any other alternative will be welcomed. 

/Frank
Reply | Threaded
Open this post in threaded view
|

Re: Send VM Metrics directly to Grafana

Gerhard Lazu
Hi,

Question: is there a way to send the VM metrics directly to Grafana?

No. You need to use one of the supported data sources (Prometheus is just one of them): https://grafana.com/docs/features/datasources/

Would Pushgateway work in your case? https://prometheus.io/docs/practices/pushing/


Linking just in case you haven't seen https://opentelemetry.io/about/ & https://opencensus.io/exporters/. Notice that Datadog accepts both metrics & traces. Maybe that would be a low-friction alternative.

On Mon, Oct 28, 2019 at 12:04 PM Frank Muller <[hidden email]> wrote:
Hi guys

While deploying one of our Erlang products in a highly secure environment (private bank), this partner required from us to export the VM metrics to Grafana. 

I know that a Prometheus Erlang client do exist, but our partner refused to deploy Prometheus (didn’t pass their pentesting phase).

Any other alternative will be welcomed. 

/Frank
Reply | Threaded
Open this post in threaded view
|

Re: Send VM Metrics directly to Grafana

Frank Muller
Thanks, I'll check them out.

Le lun. 28 oct. 2019 à 13:22, Gerhard Lazu <[hidden email]> a écrit :
Hi,

Question: is there a way to send the VM metrics directly to Grafana?

No. You need to use one of the supported data sources (Prometheus is just one of them): https://grafana.com/docs/features/datasources/

Would Pushgateway work in your case? https://prometheus.io/docs/practices/pushing/


Linking just in case you haven't seen https://opentelemetry.io/about/ & https://opencensus.io/exporters/. Notice that Datadog accepts both metrics & traces. Maybe that would be a low-friction alternative.

On Mon, Oct 28, 2019 at 12:04 PM Frank Muller <[hidden email]> wrote:
Hi guys

While deploying one of our Erlang products in a highly secure environment (private bank), this partner required from us to export the VM metrics to Grafana. 

I know that a Prometheus Erlang client do exist, but our partner refused to deploy Prometheus (didn’t pass their pentesting phase).

Any other alternative will be welcomed. 

/Frank
Reply | Threaded
Open this post in threaded view
|

Re: Send VM Metrics directly to Grafana

Edmond Begumisa

There's a useful little MIT-licensed plugin which allows Grafana to  
directly query JSON data sources over HTTP(S):

     https://grafana.com/grafana/plugins/simpod-json-datasource

To use, you'd implement the plugin's simple API on the Erlang side.

    https://github.com/simPod/grafana-json-datasource#api

If for some reason that would fail pen-testing, you could use that code as  
an example of how to implement your own Grafana datasource. But that's  
probably overkill.

- Edmond -

On Mon, 28 Oct 2019 22:35:52 +1000, Frank Muller  
<[hidden email]> wrote:

> Thanks, I'll check them out.
>
> Le lun. 28 oct. 2019 à 13:22, Gerhard Lazu <[hidden email]> a écrit :
>> Hi,
>>
>>> Question: is there a way to send the VM metrics directly to Grafana?
>>
>> No. You need to use one of the supported data sources (Prometheus is  
>> just one of them): https://>>grafana.com/docs/features/datasources/
>>
>> Would Pushgateway work in your case?  
>> https://prometheus.io/docs/practices/pushing/
>>
>> Influxdb would be my next choice:  
>> https://grafana.com/docs/features/datasources/influxdb/
>>
>> Linking just in case you haven't seen https://opentelemetry.io/about/ &  
>> https://opencensus.io/>>exporters/. Notice that Datadog accepts both  
>> metrics & traces. Maybe that would be a low-friction >>alternative.
>>
>> On Mon, Oct 28, 2019 at 12:04 PM Frank Muller  
>> <[hidden email]> wrote:
>>> Hi guys
>>>
>>> While deploying one of our Erlang products in a highly secure  
>>> environment (private bank), this >>>partner required from us to export  
>>> the VM metrics to Grafana.
>>> I know that a Prometheus Erlang client do exist, but our partner  
>>> refused to deploy Prometheus >>>(didn’t pass their pentesting phase).
>>>
>>> Any other alternative will be welcomed.
>>> /Frank
>>>
>>
>



--
Using Opera's mail client: http://www.opera.com/mail/