Note that the last specified target contains the full url instead of host and port.
This is an extension supported by VictoriaMetrics and [vmagent](https://docs.victoriametrics.com/vmagent.html) - you can use both `host:port`
and full urls in scrape target lists.
Send `SIGHUP` signal `victoria-metrics-prod` process, so it [reloads the updated `scrape.yaml`](https://docs.victoriametrics.com/vmagent.html#configuration-update):
Now the `http://localhost:8428/targets` page must contain two targets - `http://localhost:9100/metrics` and `http://localhost:8428/metrics`.
The last one should have `state: up`, since this is VictoriaMetrics itself.
Let's query the scraped metrics. Open `http://localhost:8428/vmui/` aka [vmui](https://docs.victoriametrics.com/#vmui), enter `up` in the query input field
and press `enter`. You'll see a graph for `up` metrics. It must contain two lines for the targets defined in `scrape.yaml` file above.
See [these docs](https://docs.victoriametrics.com/vmagent/#automatically-generated-metrics) about `up` metric. You can explore other scraped metrics
in `vmui` via [Prometheus metrics explorer](https://docs.victoriametrics.com/#metrics-explorer).
Let's look closely to the contents of the `scrape.yaml` file created above:
The [`scrape_configs`](https://docs.victoriametrics.com/sd_configs.html#scrape_configs) section contains a list of scrape configs.
Our `scrape.yaml` file contains two scrape configs - for `job_name: node-exporter` and for `job_name: victoriametrics`.
[vmagent](https://docs.victoriametrics.com/vmagent.html) and [single-node VictoriaMetrics](https://docs.victoriametrics.com/)
can efficiently process thousands of scrape configs in production.
Every scrape config in the list **must** contain `job_name` field - its' value is used as [`job`](https://prometheus.io/docs/concepts/jobs_instances/) label
in all the metrics scraped from targets defined in this scrape config.
Every scrape config must contain at least a single section from [this list](https://docs.victoriametrics.com/sd_configs.html#supported-service-discovery-configs).
Every scrape config may contain other options described [here](https://docs.victoriametrics.com/sd_configs.html#scrape_configs).
In our case only [`static_configs`](https://docs.victoriametrics.com/sd_configs.html#static_configs) sections are used.
These sections consist of a list of static configs according to [these docs](https://docs.victoriametrics.com/sd_configs.html#static_configs).
Every static config contains a list of `targets`, which need to be scraped. The target address is used as [`instance`](https://prometheus.io/docs/concepts/jobs_instances/)
label in all the metrics scraped from the target.
[vmagent](https://docs.victoriametrics.com/vmagent.html) and [single-node VictoriaMetrics](https://docs.victoriametrics.com/)
can efficiently process tens of thousands of targets in production. If you need scraping more targets,
then see [these docs](https://docs.victoriametrics.com/vmagent.html#scraping-big-number-of-targets).
Targets are scraped at `http` or `https` urls, which are formed according to [these rules](https://docs.victoriametrics.com/relabeling.html#how-to-modify-scrape-urls-in-targets).
It is possible to modify scrape urls via [relabeling](https://docs.victoriametrics.com/relabeling.html) if needed.
## File-based target discovery
It may be not so convenient updating `scrape.yaml` file with [`static_configs`](https://docs.victoriametrics.com/sd_configs.html#static_configs)
every time new scrape target is added, changed or removed. In this case [`file_sd_configs`](https://docs.victoriametrics.com/sd_configs.html#file_sd_configs)
can come to rescue. It allows defining a list of scrape targets in `JSON` files, and automatically updating the list of scrape targets
at [vmagent](https://docs.victoriametrics.com/vmagent.html) or [single-node VictoriaMetrics](https://docs.victoriametrics.com/) side
when the corresponding `JSON` files are updated.
Let's create `node_exporter_targets.json` file with the following conents:
```json
[
{
"targets": ["host1:9100", "host2:9100"]
}
]
```
Then create `scrape.yaml` file with the following contents:
```yaml
scrape_configs:
- job_name: node-exporter
file_sd_configs:
- files:
- node_exporter_targets.json
```
Then start [single-node VictoriaMetrics](https://docs.victoriametrics.com/) according to [these docs](https://docs.victoriametrics.com/#how-to-scrape-prometheus-exporters-such-as-node-exporter):
See [these docs](https://docs.victoriametrics.com/sd_configs.html#file_sd_configs) for details.
[vmagent](https://docs.victoriametrics.com/vmagent.html) and [single-node VictoriaMetrics](https://docs.victoriametrics.com/)
can efficiently scrape tens of thousands of scrape targets. If you need scraping more targets,
then see [these docs](https://docs.victoriametrics.com/vmagent/#scraping-big-number-of-targets).
Targets are scraped at `http` or `https` urls, which are formed according to [these rules](https://docs.victoriametrics.com/relabeling.html#how-to-modify-scrape-urls-in-targets).
It is possible to modify scrape urls via [relabeling](https://docs.victoriametrics.com/relabeling.html) if needed.
## HTTP-based target discovery
It may not so convenient maintaining a list of local files for [`file_sd_configs`](https://docs.victoriametrics.com/sd_configs.html#file_sd_configs).
In this case [`http_sd_configs`](https://docs.victoriametrics.com/sd_configs/#http_sd_configs) can help.
They allow specifying a list of `http` or `https` urls, which return targets, which need to be scraped.
For example, the following [`-promscrape.config`](https://docs.victoriametrics.com/#how-to-scrape-prometheus-exporters-such-as-node-exporter)
periodically fetches the list of targets from the specified url:
### Discovering and scraping metrics for a particular container in Kubernetes
The following [`-promscrape.config`](https://docs.victoriametrics.com/#how-to-scrape-prometheus-exporters-such-as-node-exporter)
instructs discovering and scraping metrics for all the containers with the name `my-super-app`.
It is expected that these containers expose only a single TCP port, which serves its metrics at `/metrics` page
according to [Prometheus text exposition format](https://github.com/prometheus/docs/blob/master/content/docs/instrumenting/exposition_formats.md#text-based-format):
```yaml
scrape_configs:
- job_name: my-super-app
kubernetes_sd_configs:
- role: pod
relabel_configs:
# Leave only targets with the container name, which matches the `job_name` specified above
# See https://docs.victoriametrics.com/relabeling/#how-to-modify-instance-and-job for details on `job` label.
