VICIfast
ENES
Guides & tutorials

A bigger server or a cluster? How to choose for VICIdial

More cores fix a floor that runs out of processor. They don't fix shared cores, a single point of failure or locked tables. Here's how to tell which you have.

VICIfast Support
··3 min read
Servers in a data centre — illustration for: A bigger server or a cluster? How to choose for VICIdial
Photo by Brett Sayles on Pexels

Your VICIdial server is struggling, and you can spend money on it in two ways: swap it for a bigger machine, or split the work across several. Pick the wrong one and you pay for new hardware while the real problem stays put.

The symptoms usually tell you which you need. Start with what each option can and can't do.

When a bigger server is the right fix

If the processor is what's running out, more cores work. One floor we measured ran 18 agents at an Auto dial level (lines dialed per waiting agent) of 4 to 8 on a 4-core server. CPU pressure, the share of time work sat waiting for a free core, was 25%. On 8 cores it fell to 6%.

Moving up a size keeps everything else the same: one machine, one setup, more room. If you run in the cloud, our notes on scaling a cloud VICIdial box up cover what changes during a resize.

There is a ceiling. By our estimates one machine carries about 600 Concurrent calls (calls up at the same moment) with room to spare. If your busiest hour sits well below that and the processor graph is the one in the red, a bigger single server is usually the cheaper answer.

What only a cluster fixes

Some problems come from having one machine, not a small one. No upgrade touches them.

Everything shares the same cores. On one box, Asterisk (the phone engine) and MariaDB (the database) run side by side. A busy report or a big list load takes processor time the calls needed, so the calls slow down. More cores make it rarer, but the report and the calls still share one machine. A cluster gives each its own.

One box is one failure. If that machine stops, every agent stops with it, and a bigger machine is still one machine. With two or more dialers and a standby database, a failed machine stops part of the floor, not all of it.

Your peak is heading past 600 calls. Beyond that, calls belong on dialers of their own. Our post on how many dialers a cluster needs shows how to count them.

When neither one helps

Here is the expensive mistake. Agents complain that a Disposition (the status they pick after a call) takes seconds to save, the server gets replaced, and nothing changes.

VICIdial keeps its call records in MyISAM tables, which lock the whole table while a read runs. A report that reads the call log for a few seconds makes every disposition wait behind it, and no amount of processor speeds up a wait.

We tested this. Replaying one floor's real database traffic at seven times its load, the database never needed more than 3.1 cores. What slowed first was table locks on the call log. The full numbers are in our 7x database replay.

If the screens hang but the CPU graph is calm, look for what holds the call tables before you buy anything. Keeping the call log small with VICIdial's log archiving often does more than a bigger database machine.

Deciding, in order

Ask the cheap question first: is it locks? After that, a bigger server only wins when the processor is the problem, your peak sits well short of 600 calls, one failure is something you can live with, and reports don't slow the calls.

flowchart TD
  A[Floor feels slow] --> B{Is the CPU busy?}
  B -->|No| C[Check table locks and log size first]
  B -->|Yes| D{Peak near 600 calls?}
  D -->|Yes| E[Cluster]
  D -->|No| F{Would one failed box be acceptable?}
  F -->|No| E
  F -->|Yes| G{Do reports slow the calls?}
  G -->|Yes| E
  G -->|No| H[Bigger server]

If you're weighing more signs than these, our list of signs you've outgrown one server goes through them one at a time. Our full guide to VICIdial clusters covers the layout you'd move to.

If a bigger single server is your answer, our plans list the sizes. If the chart lands on a cluster, we build it beside your current server and move the data across on a date you choose. Size your cluster and see today's prices.

Frequently asked

› Is a cluster always faster than one big server?
Not by itself. A cluster separates the calls from the database and removes the single point of failure. If the slowdown comes from locked tables, it follows you to any layout until the cause is fixed.
› Can I start with a bigger server and move to a cluster later?
Yes. Most floors start on one server and grow into a cluster. We build the cluster beside the old server and move the data across on a date you choose.

About VICIfast

We run VICIdial servers for call centres, so you don’t have to. You get the dialer set up, secured and kept running — we handle the server, the updates and the backups. Bring your own phone carrier at no markup, or use VICIfast Voice, our own US calling and numbers, from $0.0085 a minute. From $49 a month per server.

Questions? Call +1 636 556 0022 and talk to someone who runs dialers.

Citing this article

VICIfast Engineering. “A bigger server or a cluster? How to choose for VICIdial”. VICIfast LLC, November 20, 2026. Retrieved from https://vicifast.com/blog/vicidial-bigger-server-or-cluster

Have questions?

Related posts

You might be interested in

Comments

Comments are reviewed before they appear. We never publish your email.

No comments yet - be the first.