Live server capacity
Resources
What the server has available, what is using it now, and how project workloads share CPU and memory.
Snapshot 2026-09-10T21:22:20.482505Z · 0.31s sample · server refresh every 5 minObserved use, not a reservation
01
Host capacity
Short-sample CPU and current memory pressure.
CPU100.00% busy · 0.00% available
RAM58.81% used · 1.6 GiB available
Swap65.06% used
02
Usage by workload
ML, tournament bots, backdata, network, builds, podcast, agents, platform, and uncategorized work.
| Workload | CPU use | Total CPU capacity | Resident RAM | Host RAM | Processes |
|---|---|---|---|---|---|
| ML training | 0.00% | 0.00% | 0 B | 0.00% | 0 |
| Bot tournament | 174.76% | 87.38% | 1.1 GiB | 27.47% | 1 |
| Backdata | 0.00% | 0.00% | 0 B | 0.00% | 0 |
| Network & market data | 0.00% | 0.00% | 703.1 MiB | 17.96% | 1 |
| Builds | 0.00% | 0.00% | 0 B | 0.00% | 0 |
| Podcast | 0.00% | 0.00% | 0 B | 0.00% | 0 |
| Research agents | 0.00% | 0.00% | 0 B | 0.00% | 0 |
| Server platform | 3.24% | 1.62% | 125.2 MiB | 3.20% | 14 |
| Other | 9.71% | 4.86% | 179.9 MiB | 4.59% | 116 |
03
Top processes
Ranked by sampled CPU, then resident memory. Arguments and command lines are intentionally excluded.
| Process | Scope | Workload | CPU use | Resident RAM |
|---|---|---|---|---|
| cryptokingpin-t | Managed workload | tournament | 174.76% | 1.1 GiB |
| python3 | Managed workload | other | 6.47% | 12.4 MiB |
| fail2ban-server | Managed workload | platform | 3.24% | 28.4 MiB |
| kworker/u4:5-events_power_efficient | Managed workload | other | 3.24% | 0 B |
| freqtrade | Managed workload | network | 0.00% | 703.1 MiB |
| systemd-journal | Managed workload | platform | 0.00% | 38.1 MiB |
| rclone | Managed workload | other | 0.00% | 32.8 MiB |
| dockerd | Managed workload | other | 0.00% | 29.7 MiB |
| python3 | Managed workload | other | 0.00% | 25.0 MiB |
| containerd | Managed workload | other | 0.00% | 17.3 MiB |
| snapd | Managed workload | other | 0.00% | 11.8 MiB |
| systemd | Managed workload | platform | 0.00% | 9.2 MiB |
| nginx | Managed workload | platform | 0.00% | 9.2 MiB |
| nginx | Managed workload | platform | 0.00% | 9.2 MiB |
| fwupd | Managed workload | other | 0.00% | 8.1 MiB |
| docker | Managed workload | other | 0.00% | 7.4 MiB |
| nginx | Managed workload | platform | 0.00% | 6.5 MiB |
| containerd-shim | Managed workload | other | 0.00% | 4.3 MiB |
| systemd-resolve | Managed workload | platform | 0.00% | 4.2 MiB |
| systemd-network | Managed workload | platform | 0.00% | 4.0 MiB |
| udisksd | Managed workload | other | 0.00% | 4.0 MiB |
| containerd-shim | Managed workload | other | 0.00% | 4.0 MiB |
| systemd-logind | Managed workload | platform | 0.00% | 3.9 MiB |
| bash | Managed workload | other | 0.00% | 3.6 MiB |
04
How to read this page
CPU
One fully occupied core is 100% process CPU. “Total CPU capacity” divides workload CPU by installed core count. Host busy is measured independently from `/proc/stat`.
Memory
Resident memory is a point-in-time RSS total. Shared pages can be counted in multiple processes, so grouped RSS is diagnostic rather than an accounting ledger.
Freshness
The dashboard service samples live Linux `/proc` data and publishes a new public-safe snapshot every five minutes. This page checks for a newer snapshot every 30 seconds.
Privacy
Only aggregate workload labels, sanitized process names, and bounded capacity metrics are public. Exact service identifiers, command arguments, prompts, filesystem paths, raw logs, keys, and credentials never enter this dataset.