Testing SCP-Exporter
  

EduardoOliveira Lv1Posted 2026-Aug-08 01:26

Last edited by EduardoOliveira 2026-Aug-08 02:25.



Hey everyone,
I've been testing the SCP exporter using Prometheus/Grafana.
I packaged the scp-exporter into Docker (Sangfor only shares the service binary) and built a full stack around it — exporter + Prometheus + Grafana.
A few things about it:

  • Built the Docker image based on the standard scp-exporter systemd service — same config path and startup flags as the normal host install, just containerized instead of running directly on the OS

  • The exporter runs in a minimal Docker image (built on scratch, no OS layer since the binary's static — ends up around 6MB)

  • docker-compose.yml spins up all three services together, so it's basically one command to get monitoring running

  • Includes a Grafana dashboard that loads automatically (no manual setup) with:



      Host CPU/memory
      Storage capacity/IOPS/throughput
      VM stats (power state, top CPU/memory consumers)

  • Had to tune the Prometheus scrape timeout — when the SCP API is slow or unreachable, the exporter can take up to ~45s to respond, which was longer than Prometheus's default 10s timeout and was causing the target to flap

To try it:

bash
git clone https://github.com/eduardodeoliveira1995/sangforhcimonitoringstack.git
cd sangforhcimonitoringstack
cp config.example.yaml config.yaml   # fill in your SCP endpoint + access/secret key
cp .env.example .env                 # set a real Grafana admin password
docker compose up -d

Then just open localhost:3000 — the dashboard's already there under SCP Exporter.


Exporter image on Docker Hub: eddossan/scp-exporter:1.0.0

Config with real credentials never goes into the image or the repo — it's mounted in at runtime and gitignored.

Happy to answer questions or take PRs if anyone wants to extend it

559336a761523767fb.png (175.77 KB, Downloads: 22)

559336a761523767fb.png

Like this topic? Like it or reward the author.

Creating a topic earns you 5 coins. A featured or excellent topic earns you more coins. What is Coin?

Enter your mobile phone number and company name for better service. Go

Newbie585065 Lv2Posted 2026-Aug-09 17:47
  
Thanks for sharing!
Newbie167857 Lv1Posted 2026-Aug-08 13:18
  
Thank you for sharing!
AR Lv3Posted 2026-Aug-08 12:07
  
Amazing Work!!
Prosi Lv4Posted 2026-Aug-08 10:43
  
Hi, Monitoring is not just about collecting metrics.
The behavior of the monitoring chain itself—API response times, exporter processing times, scrape intervals, and timeouts—also requires attention.
Otherwise, a perfectly healthy environment might appear unstable simply because the monitoring system stops waiting too soon.

I am continuing to test this configuration and validate the generated metrics against the actual SCP/HCI environment.