The Site Issues panel sits in the summary column of the Status Page (where "Miner Control" used to be). It answers one question: "where should I go, and what's wrong there?" Instead of piecing that together from a stream of individual alerts, it condenses everything the system knows into a short, per-site list of what's actually broken and worth a look. The old alerts still fire and are the historical record — this is the at-a-glance triage view on top of them.
Collapsed, it's one line per site with a count — the glance. Sites with nothing wrong show ✓ clear. Expand a site to see its categories, expand a category to see the specific pods or gens, and expand those to see the exact problem lines. It refreshes on its own about every 5 minutes.
Everything expandable is native — no risk in clicking around. The counts and the site order just help you spot the worst site fastest; there's no "severity score," because a field guy reading it knows better than a formula whether 1 pod down beats 10 miners underperforming.
| Category | What it means |
|---|---|
| Pods Down | A whole pod is offline (lost power or network). It sweeps up its group's gen faults into the same incident — because the gens and the pod share a generator group, one outage shows as one entry instead of a dozen separate gen alarms. If an entire site goes dark it collapses further to a single "Whole site down" line. |
| Gens Down | How many generators are down — just the count. The why (the fault code, when it tripped) is already on the Status Page, so this stays a number. Gens you've disabled or tagged for maintenance are not counted. |
| Gen Issues | Running gens throwing a warning worth attention, grouped by gen group with each gen listed once and all its issues under it. Also includes bridge down — a gen that should be streaming live data but its wifi bridge went quiet. |
| Miner Issues | Miner-level trouble, per pod, matched one-to-one with the sitemap's colors: offline, zero-hash, and low-hash / underperforming. |
These are the actual criteria the panel uses. They're intentionally loose — the goal is a short list you'll trust, not every little thing.
| Flag | Fires when… |
|---|---|
| Miners offline | More than 25 offline on an online pod. A ~24-port network switch dropping looks exactly like this; the couple-always-dead miners in every pod don't reach it. So an offline cluster usually means a switch or breaker. |
| Miners zero-hash | 5+ miners online but producing nothing (dead but still responding). |
| Miners low-hash | 5+ miners running below 150 TH/s (underperforming — the sitemap paints these yellow). The "X overheating" note says how many of them have a heat code (vs. degraded hashboards or a stuck power limit). |
| Gen coolant | A gen's coolant hits 215°F+ — above the normal summer running band and closing on the 220°F shutdown (the alarm is 210°F). Below that, the auto load-shed handles the climb, so we stay quiet. |
| Gen bridge down | A gen that pushed live telemetry within the last few days has gone quiet for over 15 minutes — its wifi bridge dropped, so we've lost live data on it. |
| Gen warnings | Live oil / voltage / NG-pressure / other alarms on a running gen (heat and the chronic ones handled separately — see below). |
| Peplink / Starlink | A site or group's router is offline, or it's fallen back to cell data. |
Half the value is what it doesn't show. These are the deliberate silences — the operator knowledge baked in so the list stays short and trustworthy.
| Rule | Why |
|---|---|
| 2-minute confirm (debounce) | An issue has to stick around about 2 minutes before it appears, and it clears the instant it's gone. A 6-second network blip never shows up; a real problem shows on the next refresh. This is why a brand-new issue can take a few minutes to appear. |
| Disabled / maintenance gens | If you've disabled a gen, or tagged it Swap / Inspect / Parts / OOS, it's "already handled" — it won't nag you about a gen you're already dealing with. |
| Group attribution | When a pod loses power, the gens in its group throwing faults get folded into that one outage instead of listed as separate problems. |
| Chronic gen warnings | Some warnings (MSC link, ECU malfunction, EGT sensor faults) fire constantly fleet-wide and aren't actionable — muted so they don't drown out real ones. |
| Warm-but-full miners | A miner with a heat code that's still hashing at full rate isn't costing anything (it's green on the sitemap), so it isn't flagged. We only care about heat when it's actually reducing hashrate. |
| Non-operational sites | Out of Service, Spares, and TBD sites are inventory buckets, not live ops — hidden entirely. |
| Muted sites | Known exceptions get parked at the bottom with a reason (e.g. a site being relocated), so a situation you already know about doesn't clutter the list. |
The three miner flags use the same lines the sitemap uses to color a square, so whatever the panel names, you can go find it on the sitemap by color:
So "Stout 1: 31 low-hash" means you'll find 31 yellow squares in Stout 1 on the sitemap — the panel tells you where to go, the sitemap shows you which machines.
The alert feeds still fire and remain the full history. Site Issues is the short, deduplicated "where to go" lens on top — use it to decide where to head, then use the Status Page and sitemap for the detail.
Read How-To: Using the Status Page for the dashboard the panel lives on, and How-To: Power Management for the AS/AW/EMS toggles behind a lot of the sleep/wake behavior.