Search K
Appearance
Hardware Readiness ​
The Hardware Readiness check inspects the automation server's CPU, memory and disk, gives an overall rating, and recommends how many automation tasks to run in parallel on that server. This page explains what it measures, the calculation behind the recommendation, and the specifications we recommend for MinuteView Server and the Automations engine.
What it measures ​
| Measurement | What it is |
|---|---|
| Logical cores | The number of logical CPU processors available to the server. |
| RAM | Total memory available to the engine, in GB. |
| Free disk | Free space on the drive that hosts the engine and its working files. |
The card shows a one-line summary, for example:
8 logical cores · 16.0 GB RAM · 220 GB free on C: · Recommended: up to 8 parallel tasks
The parallel-task recommendation ​
Most automation tasks are I/O-bound — they spend their time waiting on API calls, file uploads and downloads. But some are CPU- and memory-heavy, such as DWG export, PDF generation and image processing. To stay safe for the heavy tasks, the recommendation uses logical cores as a hard ceiling rather than multiplying them, and budgets a conservative amount of memory per task.
The recommended number of parallel tasks is the smaller of two limits, then clamped to a sensible range:
byCores = logical cores
byRam = RAM (GB) ÷ 1.5 (≈ 1.5 GB budgeted per in-flight task)
recommended = min(byCores, byRam)
recommended = clamp(recommended, 2, 32)In words:
- CPU ceiling — never recommend more parallel tasks than there are logical cores, because that is the real limit on simultaneous work.
- Memory budget — allow roughly 1.5 GB per concurrent task. This is deliberately conservative so that a batch of heavy tasks (e.g. several DWG/PDF exports at once) does not exhaust RAM.
- Floor of 2 / ceiling of 32 — always recommend at least 2 (so even a small server does some parallel work), and never more than 32 (beyond which other factors — disk, network, the target systems — dominate).
Worked examples ​
| Logical cores | RAM | By cores | By RAM (÷1.5) | Recommended |
|---|---|---|---|---|
| 2 | 4 GB | 2 | 2 | 2 |
| 4 | 8 GB | 4 | 5 | 4 |
| 8 | 16 GB | 8 | 10 | 8 |
| 8 | 32 GB | 8 | 21 | 8 |
| 16 | 16 GB | 16 | 10 | 10 |
| 16 | 64 GB | 16 | 42 | 16 |
| 64 | 256 GB | 64 | 170 | 32 (ceiling) |
The recommendation is a safe upper bound, not a target. The engine's default concurrency is 1 worker; raise it towards the recommended figure gradually while watching memory and the job queue, especially for CAD-heavy workflows.
Ratings and thresholds ​
The card's badge reflects the most serious condition found:
| Badge | Condition |
|---|---|
| Failed | RAM below 4 GB, or free disk below 5 GB. Parallel processing is severely limited and output-producing tasks may fail. |
| Warning | RAM below 8 GB (consider reducing the parallel count), or free disk below 20 GB (monitor closely during heavy workloads). |
| Passed | Meets or exceeds the above — RAM ≥ 8 GB and ≥ 20 GB free disk. |
Recommended specifications ​
These pair with the System Requirements and are framed around running tasks in parallel:
| Component | Minimum | Recommended for parallel workloads |
|---|---|---|
| Logical cores | 4 | 8–16 (more for high parallelism) |
| RAM | 8 GB | 16–32 GB (24 GB+ for large CAD workflows) |
| Free disk | 20 GB | 50 GB+ on a fast SSD/NVMe, with a dedicated working directory |
Because the recommendation is the smaller of the CPU and memory limits, balance the two: adding cores without adding RAM (or vice versa) will not raise the recommended parallel-task count past whichever limit is lower.