marco.lucarelli/my_projects
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feature: grouping

Browse Source
This commit is contained in:
Marco Lucarelli
2026-02-02 12:55:32 +01:00
parent bdac561e86
commit 5c777483fc
3 changed files with 101 additions and 90 deletions
Download Patch File Download Diff File Expand all files Collapse all files

133
python/redfish-api/README.md
View File

@@ -1,47 +1,31 @@
# Redfish Exporter
A Python-based Prometheus exporter for collecting power data (Watts, Volts, Amperes) from bare metal servers using the Redfish API. This tool supports multiple vendors (e.g., HPE, Supermicro).
# Redfish-Exporter
A Python-based Prometheus exporter for collecting power data (Watts, Volts, Amperes) from bare metal servers using the Redfish API. This tool supports multiple vendors (e.g., HPE, Supermicro) and is designed to run cross-platform on Linux and Windows.
I've createtd this python script to collect Power data to analyse Watts, Volts and Amperes. If there is a better solution, feel free to replace me.
---
## Table of Contents
- [Redfish Exporter](#redfish-exporter)
- [Table of Contents](#table-of-contents)
- [Description](#description)
- [Features](#features)
- [Usage](#usage)
- [Installation](#installation)
- [Requirements](#requirements)
- [Configuration](#configuration)
- [Basic Configuration](#basic-configuration)
- [Basic Configuration](#basic-configuration-1)
- [Container](#container)
- [Legacy Installation](#legacy-installation)
- [Python Dependencies](#python-dependencies)
- [Create user](#create-user)
- [Systemd Service](#systemd-service)
- [Testet on Hardware](#testet-on-hardware)
- [License](#license)
---
## Description
This tool collects power metrics from servers using the Redfish API and exposes them in a format compatible with Prometheus. It supports both modern and legacy Redfish API versions and handles authentication for different vendors.
---
I've createtd this python script to collect Power data to analyse Watts, Volts and Amperes. If there is a better solution or you want more feature, feel free to replace me or expand my prometheus exporter.
## Features
- Collects power metrics: Watts, Volts, and Amperes.
- Supports multiple vendors (HPE, Supermicro, etc.).
- Supports grouping.
- Cross-platform compatibility (Linux and Windows).
- Graceful error handling and retry logic.
- Configurable via YAML.
- Docker support.
## Metrics Overview
| Metric | Typ | Description |
| ------------------------------------ | --------- | -------------------------------------------------------------- |
| redfish_up | Gauge | Status from host (1 = reachable, 0 = not reachable). |
| redfish_psu_line_input_voltage_volts | Gauge | Voltages per powersupply (label: host, psu_serial). |
| redfish_psu_power_input_watts | Gauge | Watts per powersupply (label: host, psu_serial). |
| redfish_psu_input_amps | Gauge | Amperes per powersupply (label: host, psu_serial). |
| redfish_system_info | Info | Systeminformation (Vendor, Model, Serial, Redfish Version). |
| redfish_request_latency_seconds | Histogram | Latency (label: host). |
| redfish_errors_total | Counter | Number of errors per host and error type (label: host, error). |
## Usage
```bash
usage: redfish_exporter.py [-h] [--config CONFIG] [--port PORT] [--interval INTERVAL]
```
usage: python main.py [-h] [--config CONFIG] [--port PORT]
Redfish Prometheus Exporter
@@ -49,28 +33,26 @@ options:
-h, --help show this help message and exit
--config CONFIG Path to config file
--port PORT Override port from config file
--interval INTERVAL Override interval from config file
```
# Installation
# Install
## Requirements
Requirements:
* just (optional)
* python 3.8+
* uv
* see `pyproject.tom`
* Python 3.8+
* see `pyproject.toml`
Install the dependencies using:
Install the dependencies using `uv`:
```bash
cd /srv/redfish-exporter
uv sync
source .venv/bin/activate
uv lock --upgrade --refresh
```
## Configuration
Create a `config.yaml` file with the following structure:
Create `config.yaml` with following structure:
### Basic Configuration
```yaml
@@ -87,7 +69,7 @@ hosts:
- host4.example.net
```
### Basic Configuration
### Advanced Configuration
```yaml
---
interval: 5
@@ -95,15 +77,18 @@ port: 8000
username: user1
password: secret1
chassis: ["1"]
group: development # set default group for all hosts
hosts:
- fqdn: host1.example.net
username: user2
password: secret2
chassis: ["0"]
group: production # use group for specific host
- fqdn: host2.example.net
username: user3
password: secret3
chassis: ["1"]
group: stage
- fqdn: host3.example.net
username: user4
password: secret4
@@ -113,30 +98,43 @@ hosts:
password: secret5
```
The `port`, `interval` are optional and can be overwritten by argument. Save default values are hardcoded.
The `port`, `interval` and `interval` are optional and can be be overridden by command-line arguments. Default values are hardcoded.
### Prometheus Configuration
```
global:
scrape_interval: 15s
evaluation_interval: 15s
# Container
scrape_configs:
- job_name: "prometheus"
static_configs:
- targets: ["localhost:9090"]
- job_name: "redfish_exporter"
static_configs:
- targets: ["localhost:8000"] # Adjust to your config
metrics_path: /metrics
scrape_interval: 15s
```
# Docker / Container
To run the Redfish Exporter in a Docker container:
```
docker buildx build -t your-tag .
docker run -it --rm --name redfish_exporter_app -p 8000:8000 your-tag:latest
docker buildx build -t redfish_exporter .
docker run -it --rm --name redfish_exporter_app -p 8000:8000 redfish_exporter:latest
```
# Legacy Installation
## Python Dependencies
```bash
mkdir /srv/redfish-exporter
# or
git clone https://github.com/dasbaum-ch/redfish-exporter.git /srv/redfish-exporter
```
## Python Dependencies
```bash
cd /srv/redfish-exporter
uv sync
source .venv/bin/activate
uv lock --upgrade --refresh
uv sync --locked
```
## Create user
@@ -145,28 +143,25 @@ sudo useradd -r -s /bin/false redfish
```
## Systemd Service
1. Copy the systemd unit file:
```bash
sudo cp redfish-exporter.service /etc/systemd/system/redfish-exporter.service
```
1. Reload and start the service:
2. Reload and start the service:
```bash
sudo systemctl daemon-reload
sudo systemctl enable --now redfish-exporter.service
```
# Testet on Hardware
Here some Server's that I have successfully testet:
* Supermicro
* AS -5126GS-TNRT2
* Redfish 1.21.0
* AS -1124US-TNRP
* Redfish 1.8.0
* HPE
* ProLiant DL380 Gen10
* Redfish 1.6.0
# License
This project is licensed under the MIT License. See the LICENSE file for details.
This project is licensed under the **MIT License**. See the [LICENSE](LICENSE) file for details.
# Testet on Hardware
Here some Server's that I have successfully testet:
| Vendor | Model | Redfish Version |
| ---------- | -------------------- | --------------- |
| Supermicro | AS-5126GS-TNRT2 | 1.21.0 |
| | AS-1124US-TNRP | 1.8.0 |
| HPE | ProLiant DL380 Gen10 | 1.6.0 |

1
python/redfish-api/config.yaml.example
View File

@@ -4,6 +4,7 @@ port: 8000
username: gloabl-user
password: global-password
chassis: ["1"] # Strings, not integers!
group: production
hosts:
- fqdn: host1.example.com
username: user1

51
python/redfish-api/redfish_exporter_v9000.py
View File

@@ -56,6 +56,7 @@ class HostConfig:
username: str
password: str
chassis: list[str] | None = None
group: str = "none"
max_retries: int = 3 # 3 retires
backoff: int = 2 # wait 2 seconds
cool_down: int = 120 # seconds to wait after too many failures
@@ -92,24 +93,28 @@ REQUEST_TIME = Summary("request_processing_seconds", "Time spent processing requ
REQUEST_LATENCY = Histogram(
"redfish_request_latency_seconds", "Time for Redfish request", ["host"]
)
UP_GAUGE = Gauge("redfish_up", "Host up/down", ["host"])
UP_GAUGE = Gauge("redfish_up", "Host up/down", ["host", "group"])
ERROR_COUNTER = Counter(
"redfish_errors_total", "Total Redfish errors", ["host", "error"]
)
VOLTAGE_GAUGE = Gauge(
"redfish_psu_line_input_voltage_volts",
"Line Input Voltage per PSU",
["host", "psu_serial"],
["host", "psu_serial", "group"],
)
WATTS_GAUGE = Gauge(
"redfish_psu_power_input_watts", "Power Input Watts per PSU", ["host", "psu_serial"]
"redfish_psu_power_input_watts",
"Power Input Watts per PSU",
["host", "psu_serial", "group"],
)
AMPS_GAUGE = Gauge(
"redfish_psu_input_amps", "Current draw in Amps per PSU", ["host", "psu_serial"]
"redfish_psu_input_amps",
"Current draw in Amps per PSU",
["host", "psu_serial", "group"],
)
# set info metric
SYSTEM_INFO = Info(
"redfish_system_info", "System information (model, serial, etc.)", ["host"]
"redfish_system_info", "System information (model, serial, etc.)", ["host", "group"]
)
@@ -170,7 +175,7 @@ async def fetch_with_retry(session, host: HostConfig, url: str) -> dict | None:
logging.warning(
"Skipping %s (in cool-down until %.1f)", host.fqdn, host.next_retry_time
)
UP_GAUGE.labels(host=host.fqdn).set(0)
UP_GAUGE.labels(host=host.fqdn, group=host.group).set(0)
return None
# Probe vendor if not already known
@@ -391,7 +396,7 @@ async def get_power_data(session, host: HostConfig):
logging.warning(
"Skipping %s (in cool-down until %.1f)", host.fqdn, host.next_retry_time
)
UP_GAUGE.labels(host=host.fqdn).set(0)
UP_GAUGE.labels(host=host.fqdn, group=host.group).set(0)
return
# Start time measurement
@@ -402,17 +407,17 @@ async def get_power_data(session, host: HostConfig):
if not resources or not resources.chassis:
logging.error("Could not discover any resources for %s", host.fqdn)
host.mark_failure()
UP_GAUGE.labels(host=host.fqdn).set(0)
UP_GAUGE.labels(host=host.fqdn, group=host.group).set(0)
return
host.mark_success()
UP_GAUGE.labels(host=host.fqdn).set(1)
UP_GAUGE.labels(host=host.fqdn, group=host.group).set(1)
chassis_url = f"https://{host.fqdn}{resources.chassis}"
chassis_data = await fetch_with_retry(session, host, chassis_url)
if not chassis_data:
host.mark_failure()
UP_GAUGE.labels(host=host.fqdn).set(0)
UP_GAUGE.labels(host=host.fqdn, group=host.group).set(0)
return
for chassis_member in chassis_data.get("Members", []):
@@ -498,13 +503,17 @@ async def get_power_data(session, host: HostConfig):
def update_prometheus_metrics(host: HostConfig, metrics: PowerMetrics):
"""Update Prometheus metrics with PowerMetrics data."""
if metrics.voltage is not None and metrics.serial:
VOLTAGE_GAUGE.labels(host=host.fqdn, psu_serial=metrics.serial).set(
metrics.voltage
)
VOLTAGE_GAUGE.labels(
host=host.fqdn, psu_serial=metrics.serial, group=host.group
).set(metrics.voltage)
if metrics.watts is not None and metrics.serial:
WATTS_GAUGE.labels(host=host.fqdn, psu_serial=metrics.serial).set(metrics.watts)
WATTS_GAUGE.labels(
host=host.fqdn, psu_serial=metrics.serial, group=host.group
).set(metrics.watts)
if metrics.amps is not None and metrics.serial:
AMPS_GAUGE.labels(host=host.fqdn, psu_serial=metrics.serial).set(metrics.amps)
AMPS_GAUGE.labels(
host=host.fqdn, psu_serial=metrics.serial, group=host.group
).set(metrics.amps)
async def get_system_info(session, host: HostConfig):
@@ -547,7 +556,7 @@ async def get_system_info(session, host: HostConfig):
serial_number = system_data.get("SerialNumber")
# Hier könnte ihre Werbung stehen
SYSTEM_INFO.labels(host=host.fqdn).info(
SYSTEM_INFO.labels(host=host.fqdn, group=host.group).info(
{
"manufacturer": manufacturer,
"model": model,
@@ -587,7 +596,8 @@ async def run_exporter(config, stop_event):
port = config.get("port", 8000)
default_username = config.get("username")
default_password = config.get("password")
default_chassis = config.get("chassis")
default_chassis = config.get("chassis", "1")
default_group = config.get("group", "none")
hosts = config["hosts"]
interval = config.get("interval", 10)
@@ -604,10 +614,15 @@ async def run_exporter(config, stop_event):
username=host_entry.get("username", default_username),
password=host_entry.get("password", default_password),
chassis=host_entry.get("chassis", default_chassis),
group=host_entry.get("group", default_group),
)
else:
hc = HostConfig(
fqdn=host_entry, username=default_username, password=default_password
fqdn=host_entry,
username=default_username,
password=default_password,
chassis=default_chassis,
group=default_group,
)
host_objs.append(hc)