Bitcoin Mining Monitoring: Essential Metrics Every Miner Should Track
You have ASICs running, a pool selected, and blocks are being found. Everything seems fine... until one day your revenue drops by 30% and you have no idea why. Was it a machine failure? Overheating? A network difficulty spike? Without proper monitoring, you are flying blind.
Monitoring is not optional in Bitcoin mining. It is the difference between a profitable operation and one that bleeds money for days before anyone notices a problem. Whether you run a single S19 in your garage or manage racks of S21s in a facility, the fundamentals are the same: watch the right numbers, understand what they mean, and act fast when something goes wrong.
This guide covers every metric that matters, how to read a mining dashboard like a pro, and exactly when you should worry versus when you can relax.
The Fundamental Metrics
Hashrate: The Heartbeat of Your Operation
Hashrate is the single most important number in mining. It tells you how many hash computations your machine performs per second. But here is where most beginners get confused: not all hashrate readings mean the same thing.
Your dashboard typically shows three different hashrate values, and understanding the distinction is critical.
| Metric | Time Window | Typical Volatility | What It Tells You | |--------|-------------|-------------------|-------------------| | Real-time (RT) | Last ~10 seconds | Very high (up to +/- 50%) | Instantaneous performance snapshot | | 1-hour average (H1) | Rolling 60 minutes | Moderate (+/- 15%) | Short-term trend, useful for spotting issues | | 24-hour average (H24) | Rolling 24 hours | Low (+/- 3-5%) | True sustained performance, closest to expected revenue |
Real-time hashrate is inherently noisy. Mining is a probabilistic process: sometimes your machine finds shares quickly, sometimes it goes through a dry spell. This is completely normal. A real-time reading that bounces between 80 TH/s and 140 TH/s on a machine rated at 110 TH/s is perfectly healthy.
The H1 average smooths things out and is your best early warning indicator. If this number drops significantly below the machine's rated hashrate, something is happening right now that needs attention.
The H24 average is what your revenue actually tracks. This is the number you compare against the manufacturer's specifications. For example, if you own an Antminer S21 rated at 200 TH/s and your H24 consistently reads 185 TH/s, you are getting about 92.5% efficiency, which is acceptable. Below 80% of the rated hashrate, you have a problem worth investigating.
Chip Temperature: The Silent Killer
Heat is the number one enemy of ASIC miners. Every chip has a thermal limit, and exceeding it leads to throttling (automatic performance reduction), accelerated wear, and eventually permanent damage.
Most modern ASICs report two temperature values: the board temperature (ambient air around the circuit boards) and the chip temperature (the actual temperature of the hashing chips). The chip temperature is the one that matters most.
| Chip Temperature | Status | Action Required | |-----------------|--------|-----------------| | < 55C | Cool | None. Excellent conditions, typically seen in cold climates or with aggressive cooling. | | 55C - 70C | Optimal | None. This is the sweet spot for performance and longevity. | | 70C - 80C | Warm | Monitor closely. Check that fans are running at appropriate speed and airflow is not obstructed. | | 80C - 85C | Hot | Investigate immediately. Clean dust filters, check ambient temperature, verify fan operation. The machine may begin throttling. | | > 85C | Critical | Shut down or reduce clock speed. At this level, most ASICs engage automatic throttling, cutting your hashrate significantly. Sustained operation above 85C will shorten chip lifespan dramatically. |
Throttling deserves special attention because it is sneaky. When chips overheat, the firmware automatically reduces clock frequency to bring temperatures down. Your machine keeps running and appears healthy at a glance, but your hashrate silently drops 20-40%. Without temperature monitoring, you might not notice for days or weeks, losing revenue the entire time.
Shares: Accepted, Rejected, and Stale
Shares are proof of work submitted to your mining pool. They come in three flavors, and only one of them earns you money.
- Accepted shares: Valid work that your pool counted toward your payout. This is what you want.
- Rejected shares: Work your pool rejected because it was invalid. A small rejection rate (under 0.5%) is normal. Above 1%, you may have hardware issues or firmware problems.
- Stale shares: Work that was valid when your machine computed it, but arrived at the pool too late because the block had already changed. Stale shares earn you nothing.
Your stale rate is particularly important because it directly eats into revenue. Every stale share represents electricity you paid for but got zero return on. A stale rate under 0.5% is excellent, 0.5-1.5% is typical, and anything above 2% signals a network or configuration problem that needs fixing. For a deep dive on reducing stales, check out our guide on how to reduce your stale rate in Bitcoin mining.
Uptime
Uptime measures what percentage of the time your miner is actually running and hashing. It sounds simple, but the financial impact of downtime is significant.
At current prices (BTC around $64,000), an Antminer S21 generating roughly $15-18 per day loses about $0.60-$0.75 for every hour of downtime. That does not sound like much until you realize that a machine that reboots twice a day and takes 5 minutes each time to stabilize loses over $350 per year, just from restarts. Scale that across 50 machines and you are looking at $17,500 in lost revenue from a problem that good monitoring would catch immediately.
Target uptime should be 99%+ for any serious operation.
Estimated Revenue
Most dashboards display an estimated daily revenue based on your current hashrate, the network difficulty, and the current BTC price. This number is useful as a quick sanity check but comes with caveats.
Estimated revenue fluctuates constantly because it depends on real-time hashrate (which is noisy), transaction fees (which vary block to block), and the BTC/USD price. Do not panic if the number swings 20% within an hour. Instead, look at the trailing 7-day average for a realistic picture of what your machines are earning.
For strategies to maximize what your hardware earns, see our guide on optimizing Bitcoin mining revenue.
How to Read a Mining Dashboard
A good mining dashboard gives you an overview of your entire operation at a glance, then lets you drill down into individual machines. Here is how to read one effectively.
The Overview Panel
Start with the big picture. The overview should show your total fleet hashrate, number of active machines, overall stale rate, and estimated daily earnings. Scan these numbers first. If everything looks normal, you may not need to go further.
Hashrate Chart Patterns
The hashrate chart over time is the most information-dense element on your dashboard. Learning to read its shape will save you hours of troubleshooting.
Stable, flat line (with minor noise): This is what healthy mining looks like. The line stays close to your expected hashrate with small random fluctuations. No action needed.
Sudden drop to zero: A machine stopped. This could be a power outage, a crash, a network disconnection, or a tripped breaker. Check immediately. The longer a machine sits idle, the more revenue you lose.
Sawtooth pattern (repeated drops and recoveries): This classic pattern indicates your miner keeps disconnecting and reconnecting to the pool. Common causes include unstable internet, a flaky Ethernet cable, pool-side issues, or DNS resolution problems. Each reconnection cycle wastes time and produces stale shares. Consider using a mining proxy to stabilize the connection between your machines and the pool.
Gradual decline over hours or days: This is the hallmark of thermal throttling. As ambient temperature rises (afternoon heat, changing seasons, blocked ventilation), chip temperatures climb, and the firmware quietly reduces clock speeds to compensate. Check your temperature graphs alongside the hashrate chart; if they move in opposite directions, overheating is your culprit.
Per-Machine View
When the overview flags an anomaly, drill into the individual machine view. Here you can see specific chip temperatures, fan speeds, per-board hashrate (most ASICs have 3 hash boards), and error logs. A machine where one board shows significantly lower hashrate than the other two likely has a failing board that needs repair or replacement.
When to Worry vs. When It Is Normal
Not every fluctuation is a crisis. Here is a quick reference to help you calibrate your response.
Normal Situations (No Action Required)
| What You See | Why It Happens | |-------------|---------------| | Real-time hashrate swings +/- 30-50% | Normal variance in share finding. Probabilistic process. | | H1 hashrate dips 10-15% occasionally | Short-term luck variance. Will average out over 24 hours. | | 1-2 rejected shares per day | Minor, expected. Hardware occasionally produces invalid work. | | Stale rate of 0.3-0.8% | Normal for most pool connections. | | Revenue estimate changes hourly | BTC price, mempool fees, and difficulty all fluctuate constantly. | | Temperature rises 3-5C in the afternoon | Ambient temperature variation. Monitor but do not panic. |
Alert Situations (Investigate Immediately)
| What You See | Likely Cause | Action | |-------------|-------------|--------| | H1 hashrate drops below 80% of rated | Overheating, failing board, firmware issue | Check temperatures, inspect boards, restart if needed | | Machine offline for more than 5 minutes | Crash, power loss, network failure | Physical check, verify power and network, restart | | Chip temperature above 82C sustained | Cooling failure, dust buildup, high ambient temp | Clean filters, check fans, reduce overclock, improve ventilation | | Stale rate above 2% | Network latency, pool issues, bad configuration | Check internet connection, switch pool endpoint, consider a proxy | | Rejection rate above 1% | Hardware fault, bad firmware, overclocking too aggressive | Reduce overclock, update firmware, test with stock settings | | One hash board significantly underperforming | Failing board, loose connector | Reseat connectors, test board independently, RMA if needed | | Fan speed at 100% with rising temperatures | Blocked airflow, failed fan, extreme ambient heat | Immediate inspection of airflow path, check for obstructions |
Setting Up Effective Monitoring
There are several approaches to monitoring, ranging from simple to comprehensive.
Option 1: Manual Checks
The simplest method: log into each machine's web interface daily, check the numbers, and note anything unusual. This works for 1-3 machines but becomes tedious and unreliable beyond that. You will inevitably miss things, especially overnight failures.
Pros: No setup required, no additional cost. Cons: Does not scale, no alerting, relies on you remembering to check.
Option 2: Pool Dashboard
Every major mining pool (Braiins Pool, F2Pool, Antpool, ViaBTC, etc.) provides a web dashboard showing your workers, their hashrates, and your earnings. This is a solid step up from manual checks because the pool monitors your machines 24/7 and many pools offer basic email alerts for offline workers.
Pros: Free, always available, includes earnings data. Cons: Limited to pool-visible data (no temperature, fan speed, or on-device metrics), alert options are usually basic, and data resolution is lower (typically 10-minute or 1-hour averages only).
Option 3: Mining Proxy with Integrated Monitoring
A mining proxy sits between your ASICs and the pool, routing all traffic through a central point. Advanced proxies like the one offered by Minent provide real-time dashboards with full visibility into hashrate, stale rates, per-machine performance, and alerting, all without needing to access each machine individually.
This approach is the gold standard for operations of any size because the proxy sees everything: every share submitted, every rejection, every reconnection. It can alert you within minutes of a problem and provides the data you need to diagnose issues quickly.
Pros: Comprehensive monitoring, real-time alerts, reduces stale rates through optimized pool connections, single pane of glass for your entire fleet. Cons: Requires initial setup (though typically straightforward).
To learn more about how a proxy improves your mining operation, read our complete guide to Bitcoin mining proxies. You can also try Minent's proxy for free or request a demo to see the monitoring dashboard in action.
The 3 Essential Alerts Every Miner Needs
If you set up nothing else, configure these three alerts. They cover the vast majority of revenue-threatening situations.
1. Machine Offline for More Than 5 Minutes
This is the most critical alert. A machine that stops hashing earns zero revenue while still potentially consuming standby power. Five minutes gives enough buffer to avoid false alarms from brief restarts, while catching genuine failures before they cost you real money.
At BTC ~$64,000, an S21 generates roughly $0.012 per minute. A 12-hour undetected outage costs about $8-9 per machine. Across a fleet, overnight failures can easily cost hundreds of dollars before morning.
2. Hashrate Drops Below 80% of Rated
If a machine's H1 average falls below 80% of its expected hashrate, something is wrong. This threshold filters out normal variance while catching throttling, failing boards, and firmware issues. Thermal throttling alone can silently reduce your revenue by 20-40% if it goes undetected.
3. Chip Temperature Exceeds 82C
This alert catches overheating before it triggers heavy throttling at 85C+. At 82C, you still have time to act: clean filters, adjust fan curves, improve ventilation, or reduce overclocking. Waiting until 85C means the machine has already started cutting its own performance and you are losing revenue every minute.
Putting It All Together
Monitoring is not about obsessively watching numbers all day. It is about building a system that watches for you and only demands your attention when something genuinely needs it.
Start by understanding the metrics: what hashrate, temperature, shares, and uptime actually mean and what ranges are healthy. Learn to read your dashboard's hashrate chart, because those visual patterns tell you more at a glance than any single number. Set up the three essential alerts so you are notified of problems whether you are at your desk or asleep.
And if you want monitoring that works out of the box with zero configuration headaches, Minent's mining proxy gives you a real-time dashboard, intelligent alerts, and stale rate optimization in a single tool. Sign up for free or book a demo to see how it works with your setup.
Your machines run 24/7. Your monitoring should too.