Looking for some help on a project/experiment.
This will involve 10x storj nodes.
5x will be hosted locally, sharing the same public IP.
5x will be hosted remotely, each with unique locations and public IPs.
Not within the same public /24 subnet.
These will all be turned on within a reasonable short time of each other.
Odriod-HC2 with 6tb to 12tb drives to each node. Where I need help is not well versed in Linux or Docker. Metrics and analytics are a requirement: ping, cpu load, memory usage, bandwidth usage, I/O performance, disc volume tracking. To compare 5x nodes centralized vs 5x decentralized.
Have a server with EXSi and static IPs locally for a centralized motoring system that everything can be connected back for reporting.
The monitoring system and equipment is etched in pencil. Seen grafana + influxDB + prometheus monitoring on the forums. I’m all ears on other options to accomplish the same goal. Even if there is a cheaper or more solid solutions than the Odriod-HC2. Granted these have an output of 25w.
Which is around $24.28/per node/per year depending on the kWh charge in your area.
Calculation above is based off: $00.115 per kWh charge.
- 10x nodes [5x local, 5x remote]
- Remote access to remote nodes
- Monitoring on local and remote nodes: ping, cpu load, memory usage, bandwidth usage, I/O performance, disc volume tracking
Will publicize the findings from this project/experiment as an on going blog.
Anyone up for helping me turn this idea into a reality?