Hey Brian, Actually to my knowledge, Datadog can also pull metrics from the
endpoints. Thus, I want to give it an endpoint which has all the filtered
metrics I want to send to datadog. That's why I use /federate endpoint
taking data from my own Prometheus server. Is there any easy way to send
metrics to datadog without using a federate endpoint?
On Sunday, March 1, 2020 at 3:25:59 AM UTC+5:30, Brian Candler wrote:
>
> Maybe you've misunderstood scraping and federation.
>
> Prometheus scrapes are "pulls". When a prometheus scrape job runs, it
> makes an outbound HTTP connection to a remote server (an "exporter") and
> reads back metrics from it. So in your example here:
>
> 31 - job_name: 'prom'$
> 32 metrics_path: '/'$
> 33 static_configs:$
> 34 - targets: ['localhost:12156']$
>
> prometheus periodically connects to localhost port 12156, and whatever is
> listening on that port is a prometheus "exporter" which responds with
> metrics. You can also talk to it like this:
>
> curl localhost:12156
>
> and that server will respond with lines of the form
>
> metricname{label="value",...} value
>
> These are what prometheus ingests into its timeseries database. No data
> is sent *from* prometheus to the exporter.
>
> Federation is a special case of scraping: it's when one prometheus server
> scrapes the /federate endpoint of another prometheus server, to read out
> metrics which that server has already collected. (For example: a central
> aggregation server is scraping the /federate endpoint of remote servers in
> other data centres).
>
> Therefore, you would only configure your own prometheus server to scrape a
> /federate endpoint when you need to scrape metrics out of another, remote
> prometheus server.
>
> So this makes no sense:
>
> 24 - job_name: 'federate'$
> 25 metrics_path: '/federate'$
> 26 params:~$
> 27 'match[]':$
> 28 - '{job="prom"}'$
> 29 static_configs:$
> 30 - targets: ['localhost:9090']$
>
> Here you're trying to scrape your own /federate endpoint, which means
> collecting data from yourself and writing it back to yourself.
>
> The original question was: how do you get data from prometheus to
> datadog? I don't know: it depends on what interface(s) datadog provides
> for collecting/ingesting data, and datadog is a commercial cloud-hosted
> service. (Perhaps their technical support can help?)
>
> However you said originally: *"I know that we need to make changes in
> datadog's conf file present in openmetrics.d directory. But as Prometheus
> only shows metrics about it's own server on /metrics. How can we send them
> to datadog?"*
>
> If datadog's agent is able to scrape an exporter, then point it at your
> /federate?match=... endpoint rather than your /metrics endpoint.
>
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/9cfae09d-b3e8-4e54-afbd-df3773accd40%40googlegroups.com.
