How Smart Job Routing Works in a 3D Print Farm
When you have one printer, job management is trivial: start the job, wait for it to finish, start the next one. When you have ten printers, job management is a part-time job. When you have twenty, it's a full-time one — unless you automate it.
This is what a job queue is for.
What goes wrong without a queue
On a farm without a job queue, job assignment is manual. Someone decides which printer gets which job — either by walking the floor, checking a spreadsheet, or fielding messages from operators. This creates a predictable set of problems:
Jobs wait for human attention. A printer finishes at 3am. Nobody notices until the morning shift. That's 6+ hours of idle machine time that didn't have to happen.
Wrong jobs go to wrong printers. A job that needs a 0.2mm nozzle gets sent to a printer loaded with 0.4mm, or a PETG job goes to a printer whose bed isn't clean from the last ABS print. You catch it when it fails, not before.
Operators step on each other. Two operators both see an idle printer and assign different jobs to it. One assignment gets silently dropped, or both jobs start and one gets cancelled mid-print.
No priority. Rush orders sit in the same mental queue as background production. Everything is "equally urgent" until it isn't.
A job queue solves all of these — but only if it's a smart queue that knows about your printers.
What "smart" routing actually means
A naive queue assigns the next job to the next available printer. That's better than nothing, but it ignores whether the assignment is physically valid.
Smart routing adds constraints:
Material matching. A job requiring PETG should only route to a printer that has PETG loaded. Not PLA. Not whatever was left from the last run. The queue knows what's loaded on each printer and only routes compatible jobs.
Nozzle size matching. Fine-detail jobs (0.2mm) route to printers with 0.2mm nozzles. High-throughput jobs route to 0.4mm or larger. This isn't just quality — running a fine job through a 0.6mm nozzle produces failed prints.
Printer capability matching. Not all printers in a fleet are equivalent. Some may have an AMS loaded with specific colors. Some may be reserved for particular customers or material types. The queue respects these constraints when routing.
Priority ordering. High-priority jobs (rush orders, customer SLAs) go to the next available compatible printer. Low-priority jobs fill in the gaps. You control priority at job creation; the queue handles the rest.
Capacity awareness. If three printers are compatible with a job but one has a partial filament spool that won't complete the run, the queue routes to a printer with sufficient material.
The result: jobs land on printers that can actually complete them, in priority order, without human intervention.
The idle time problem
Idle time — the gap between a print completing and the next job starting — is the biggest throughput lever most farms aren't managing.
On a 20-printer farm charging $6/hr per machine:
- 2 hours average idle per printer per day = $240/day in lost throughput
- 1 hour average idle = $120/day
- 30 minutes average idle = $60/day
Reducing idle time from 2 hours to 30 minutes recovers $180/day → ~$5,400/month. On a $49/month Pro plan, that's a 110x return.
The mechanism is simple: when a printer finishes, the queue automatically assigns the next compatible job. No operator action required. The printer is back in production within the time it takes to clear the plate and start the print — typically 5–10 minutes.
Without a queue, that same transition requires someone to notice the printer finished, find the right next job, and manually assign it. At 3am, that doesn't happen.
How the queue interacts with AMS
Bambu Lab's AMS (Automatic Material System) changes the material-matching problem. With AMS, a single printer can have up to 16 filament spools loaded simultaneously — 4 AMS units × 4 slots each.
This creates more routing flexibility (more jobs are compatible with more printers) but also more matching complexity. A job needing four specific colors can only route to a printer with all four of those colors loaded in AMS slots.
Print Hive's job router tracks AMS slot contents per printer — color, material type, brand if you track it — and matches jobs to printers that have all required slots filled. A 4-color print won't route to a printer with only 3 of the 4 required filaments.
When a filament slot runs out mid-print, the AMS prompts for a swap. The queue doesn't prevent this — it routes based on what's loaded when the job starts, not what will remain when it ends. Filament inventory management (tracking spool weights, estimating remaining material) is a separate discipline that some farms automate and others manage manually.
Building a job queue workflow that actually gets used
A job queue only improves throughput if everyone uses it consistently. Farms that implement a queue and then let operators bypass it get the overhead without the benefit.
The habits that make a queue work:
All jobs enter through the queue. No direct printer assignments. Even urgent jobs go into the queue with high priority — they'll route immediately to the next available compatible printer.
Job metadata is complete at creation. Material type, nozzle size, priority, estimated filament usage. Incomplete job entries break routing. The friction of filling these in at creation is much less than the friction of a mis-routed print.
Printer profiles stay current. If you swap a 0.4mm nozzle for a 0.2mm nozzle, update the printer profile before the next job routes. Stale printer data causes mis-routes. A 30-second update when you make a hardware change prevents 3-hour failed prints.
Completed plates get cleared promptly. The queue can't start the next job if the plate isn't clear. On a farm, plate clearing is the physical bottleneck. Some farms assign a dedicated operator to this; others route it based on shift schedules.
Review the queue before each shift. Spend 5 minutes at shift start reviewing what's queued, what's printing, and whether any printers need maintenance. This converts reactive fire-fighting into proactive management.
When jobs should bypass the queue
There are legitimate reasons to override the queue:
- Maintenance runs — calibration prints, first-layer tests, nozzle priming — aren't production jobs and shouldn't compete with the queue
- Customer-specific machines — if a printer is dedicated to one customer's work, that printer should be excluded from the general pool and managed separately
- In-progress manual setups — multi-part prints that require manual plate arrangement before queuing
The key distinction: intentional bypasses are fine; habitual bypasses mean the queue setup isn't working. If operators regularly go around the queue, find out why and fix it — the friction is either in job entry, routing constraints, or operator training.
Queue metrics worth tracking
A queue generates data you can use to improve operations:
Average idle time per printer. Trending down is good. Trending up means either insufficient jobs in queue or routing constraints that are too restrictive.
Routing failure rate. How often does a job sit in queue because no compatible printer is available? High failure rates indicate either missing printer capabilities or jobs with incorrect metadata.
Printer utilization by machine. If some printers consistently run at 90% and others at 60%, the queue is routing unevenly — usually a material or capability constraint issue.
Job completion rate. What fraction of queued jobs complete vs. fail or get cancelled? Improvement here comes from better printer maintenance and tuning, not queue changes — but the queue is where you measure it.
Print Hive's job queue is available on Starter ($19/mo) and above. Free tier supports manual job management for up to 2 printers. Start free →