What server does VICIdial need for 60 agents?
For 60 agents dialing predictive at three lines each, VICIdial needs Velocity, a server with 16 dedicated cores; for 60 inbound or manual agents, Ultimate with 16 shared cores is enough. VICIfast runs either one as managed VICIdial hosting: your own server with the usual VICIdial admin and agent screen, and the patching, firewall, SSL and 24/7 monitoring done for you.
| How the floor dials | Calls at peak | Smallest server that fits | Peak CPU |
|---|---|---|---|
| Outbound, predictive 3 lines per waiting agent | 180 | Velocity · 16 dedicated cores · $899/mo | 68% |
| Inbound every agent on a call, 10% more waiting | 66 | Ultimate · 16 shared cores · $749/mo | 34% |
| Manual dialing one line per agent | 60 | Ultimate · 16 shared cores · $749/mo | 32% |
- Bandwidth at peak
- 21 Mbps
- Predictive outbound, both legs
- Carrier channels
- 220
- Concurrent calls to ask your carrier for
- Recordings per day
- 18.1 GB
- 8 dialing hours
- Days of recordings on disk
- 17
- On Velocity, before offloading
The rows are the same 60 agents dialing three different ways. "Calls at peak" is the busiest minute, and "Peak CPU" is how hard that minute works the server. The tiles under the table belong to the predictive row. The agent counts on pricing are suggestions for a typical mix, which is why a predictive floor lands on a bigger plan than its headcount suggests.
Why do 60 predictive agents need dedicated cores?
Because 180 live calls at the busiest minute take a 16-core shared server past the 60% ceiling the calculator keeps on shared cores, while the same 16 cores, dedicated, run at about 68% - just under their 70% ceiling. Shared cores get the lower ceiling because a neighbour on the same host can borrow them, and the busiest minute is exactly when you cannot spare them.
60 is the crossover. Ten agents either side, the answer changes:
| Predictive agents | Live calls at peak | Smallest server | Peak CPU |
|---|---|---|---|
| 50 | 150 | Ultimate, 16 shared cores | about 56% |
| 60 | 180 | Velocity, 16 dedicated cores | about 68% |
| 70 | 210 | Turbo, 32 dedicated cores | about 40% |
The VICIdial server sizing guide explains the reasoning behind each step.
How much headroom should a 60-agent server have?
Enough that your busiest minute stays under the ceiling after the changes you already know are coming - and at 68%, Velocity has only a little room left for a 60-agent predictive floor. On a floor this size, headroom goes to:
- A higher dial level. Going from three to four lines per agent on a thin list adds a third more outbound calls.
- More agents. At 70 predictive agents the calculator moves you up to Turbo.
- Inbound on the same server. A blended floor adds its inbound calls, including callers waiting in queue, to its outbound lines.
Agent screens lagging and agents dropping out at the busiest hour are the usual signs a server is near its limit. Spotting capacity problems before they hit lists what to watch, so you see it in the numbers before your agents feel it.
When should you resize before a busy season?
Resize before the first heavy week of the season, not in the middle of it, and schedule the move for your quietest hour. One-click resize moves the same server to a bigger plan in place: your data, settings and IP address stay, and the downtime is a few minutes. You can set it to run at a future time, such as 03:00 server time, instead of doing it live.
Upgrades take effect straight away and are pro-rated. A downgrade waits for the end of the billing period, so you can size up for a season and come back down after it. Scaling VICIdial vertically in the cloud covers when a bigger server is the right fix and when the problem is somewhere else.
When don't 60 agents need dedicated cores?
If your 60 agents take inbound calls or dial by hand, Ultimate on shared cores carries them, and Velocity buys headroom you will not use. The same is true if only 50 of the 60 dial at the busiest minute: size for agents logged in at once, not the headcount.
Where to start
Enter your own floor in the capacity calculator - agents, lines per agent and inbound calls at the busiest minute - and compare the peak CPU on each plan before you buy. Then start the 7-day free trial (card required); the server is ready in under 40 seconds. If a busy season is coming and you want a second opinion on timing, ask us - a person reads every message.