I’m a solo developer in France. I built ZAMUT over the past week, a monitoring and control console specifically for Equihash mining. Before I write production code I’d rather find out from people who do this daily whether I’m solving a real problem or one I invented at my desk.
The assumption I’m working from , please tell me if it’s wrong.
My impression is that every fleet tool out there is built around GPU rigs first, with ASIC support added afterwards, and Equihash treated as one algorithm among forty. So a Z15 operator ends up with a vendor web UI per machine, a spreadsheet, or a generic tool that doesn’t really understand what it’s looking at.
I might be completely wrong. If you run Z15s and you’re happy with what you already use, that is the single most useful thing you could tell me and it saves me six months.
What I’ve built so far.
An interface prototype, not a product. It runs on simulated data. What exists:
- One view that switches between a single machine and a whole fleet , the solo case is just a fleet of one
- Per-unit telemetry: Sol/s, temperature, watts, Sol/W
- A panel that flags thermal drift before throttling starts, and rejected-share rate
- Net margin, not gross , it asks your kWh price and tells you when you’re losing money
- Pairing over LAN, with ASICs adopted directly over their API , no agent installed on the miner itself
Screenshots below.
What it does not do.
It does not mine. There is no backend. Nothing is downloadable. I’m at the stage where changing direction costs nothing, which is why I’m asking now instead of after.
Four questions.
- How many Equihash ASICs are you running right now? Rough number is fine. It changes everything about what’s worth building.
- What was the last thing that went wrong with one of your machines, and how did you find out about it? This is the one I care about most. Not what you’d like in theory , what actually happened, and how long it took you to notice.
- What do you use today to watch them? Vendor UI, HiveOS, Awesome Miner, minerstat, custom scripts, nothing. And if you’re on third-party firmware (Vnish, LuxOS, other), is the cgminer API on 4028 still reachable, or is it locked down?
- Do you pay for anything to manage your machines today? If so, what and roughly how much . I’m trying to understand what this category is actually worth to people.
What I’m not asking for.
Not money, not a launch, no token. If this becomes something the ecosystem wants there’s a grants process and I’d go through it properly. Right now I want to know whether the problem is real.
Happy to be told this is a bad idea.
2 Likes
As a former moderate gpu home miner back in the day, it looks like you’re trying to build a fleet-wide software management system like Foreman. I’d say it’d be interesting for a ASIC specific system like the Z15/Z15Pros but as a miner your really only tracking a couple of key metrics: Uptime, hashrate, Temperature, stale shares if any and latency. There are board specific metrics you might want to track to overclock a device and understand limitations of a particular board, but that would be in your management software. Tbh the graphics are cool, but most would just be looking for logging type of datasets to see trend divergence or things that might be affecting performance. Cool idea but check out foreman and what HiveOS and RaveOS does to manage workers remotely
1 Like
Useful, thanks.
The metric list is what I needed. Uptime isn’t tracked at all right now and you put it first. Stale shares and latency are in there but only as text, not as values I follow over time.
The logging point is the real one. What I built is a snapshot , everything shows “right now.” Trend divergence needs history and there is none. Same for board level: a Z15 Pro has three hashboards and I’m showing one number, so a dead board just reads as a third less hashrate.
Three questions:
-
Does Foreman actually run Z15 Pros on Equihash, or is it SHA-256 in practice? Their site doesn’t say and it’s the thing I most need to know.
-
For trend data , do the existing tools cover it, or is that where people still fall back to their own scripts?
-
What time window actually mattered for spotting divergence? Hours, days, weeks?
Thanks for your response
I’m not sure if I follow. On the ASICs they usually have their own firmware from the provider Bitmain in this case that the miner runs on. Foreman is just an aggregator and pulls measurements and other data from a group of miners (or typically called workers for each machine) and then allows you to monitor and track how each worker is doing. Some systems are designed to have 3rd party firmwares to run on ASICs like BRAIINS on Bitcoin and I’m sure there are some others as well. They’ll give you remote access to actually modify states of the workers to tweak things like overclocks, so that you don’t have to get into the actually firmware screen of each individual machine.
Trend data is typically covered in some form or fashion, either directly from the mining firmware of the miner provider or sometimes through the pool (usually limited) and through 3rd party monitoring softwares like Foreman.
Here is the site for foreman though so you can see what they do: Mining Solutions - OBM
I could be wrong, but I think ZEC and other miners are covered, but check out some mining content creators and I’m sure you can get some answer pretty easily around foreman and other like softwares
That clears it up . I was thinking of Foreman as something that runs on the miner rather than something that polls it. If it’s an aggregator reading the API, then the algorithm barely matters to it, which undercuts most of what I assumed going in.
I went through their docs , the Z-series shows up in their filtering examples, so that answers it.
Thanks for taking the time 
