Why VICIdial is not dialing: reading the hopper, in order
Twelve checks, in the order the hopper applies them, that explain why a VICIdial campaign has leads but places no calls. With the queries to run.

To the point
VICIdial dials only from the hopper, a small table of the next leads per campaign. Every minute the hopper script writes why it is or is not filling, and it applies its rules in a fixed order, so the first check that fails is your answer.
- Read the row in
vicidial_campaign_stats_debugwhereserver_ipisHOPPERfirst. Often you are done. - An expired list is as invisible as an inactive one.
- Calling hours are judged at 104 GMT offsets, in quarter-hour steps.
- A wrong server clock shifts every calling window by the difference.
Start with what the hopper says
The hopper is vicidial_hopper. So "not dialing" means either the hopper is empty or nobody is taking from it. AST_VDhopper.pl writes one row per campaign at the end of every pass:
SELECT entry_time, debug_output
FROM vicidial_campaign_stats_debug
WHERE campaign_id = 'YOURCAMP' AND server_ip = 'HOPPER';
On our test campaign it read campaign lists count ACTIVE: 0 | TOTAL: 0, then leads to call count: 0, then ERROR CANNOT ADD ANY LEADS TO HOPPER. That is the whole diagnosis.
The twelve checks, in order
- Is the dialer alive?
SELECT last_update FROM server_updater;should be seconds old, and the debugentry_timeunder a minute. If not, stop. - Is the campaign active? Read
active,dial_method,hopper_level,local_call_time,dial_statusesandlead_filter_id. - Is there a usable list? The hopper sees lists attached to the campaign with
active = 'Y'andexpiration_date >= today. - Do the lists hold leads? Count per list with
FORCE INDEX(list_id), the index the hopper forces. - Are leads in a dialed status?
dial_statusesis a space-separated string ending in-. A list worked for a month can be full and hold almost nothing dialable. - Called since the last reset? The base condition is
called_since_last_reset = 'N'with a dialed status in an active list. A lead called once stays out until the list is reset. - Inside calling hours where the lead is? The check that catches the most people. See below.
- Does a lead filter remove them? If
lead_filter_idis notNONE, its SQL is added. Count with and without it. - Attempt caps.
call_count_limitanddrop_lockout_timeact in the count. The daily and 24-hour limits act when a lead is inserted, so the count can look healthy while leads are refused. - DNC. Internal and campaign matches are dropped at insert time too. If the dialable figure is high and the hopper stays thin, measure the overlap.
- Hopper against demand. "Ready" means
READY,RHOLDorRQUEUE. The hopper refills towardhopper_levelonce a minute. - Agents. A full hopper with nobody logged in dials nothing. If agents are present and lines do not go out, see lines per agent.
Set it up, step by step
On our test server, campaign SUNQ4 and list 4001.
In VICIdial's admin

- Campaign: the list must be attached to the campaign you are checking.
- Active: must be
Y. - Expiration Date: a date before today removes the list, even when Active is
Y.

- Lead Filter: anything but
NONEadds its SQL to the count. - Drop Lockout Time: keeps leads that just dropped out of the hopper for a while.
- Call Count Limit: adds
called_count < limit. - Minimum Hopper Level: the size the hopper refills toward.
In the Trustane console

- Fix → Lists: jumps to the lists when the active-lists check fails.
- Fix → Hopper level: jumps to the hopper setting when the hopper cannot keep up with demand. The page also shows the hopper's own last run for the campaign.
Calling hours, holidays and the clock
The hopper asks what time it is for a lead at each GMT offset, from +13.00 down to -12.75 in steps of 0.25. That is 104 values. A lead is in the window when start <= HHMM < stop: start inclusive, stop not. A weekday of 0/0 falls back to the default hours, and 2400 means closed.
- Holidays narrow the day. They raise the start and lower the stop. A 2400/2400 holiday closes the day.
A box has three clocks: the OS zone, servers.local_gmt (which the hopper uses), and PHP's date.timezone. On the stock ViciBox 12 image we tested, the OS ran UTC while local_gmt said -5.00, so the hopper's local time was five hours out. On 9am to 9pm, the campaign would start and stop five hours wrong.
What we measured
One small server, read-only queries, single runs.

The scans took 1.35 s, 1.38 s and 3.4 s, cheap enough for a live box. About 5% of eligible leads were on DNC. We could not confirm our final count against dialable_leads on a live list. Two independent scans agreeing is not the same thing.
Questions people ask
The list has 40,000 NEW leads. Why is the hopper empty?
Check 6: leads called since the last reset stay out until the list is reset, whatever their status.
Why is the hopper full but nothing dials?
Check 12: nobody is logged in, or lines are capped.
Next steps
Read the HOPPER row, work down the checks and stop at the first failure. For calling hours, compare the four clocks before touching the call time. On hosted servers the console runs these checks for every campaign: order a server, migrate one or see VICIdial hosting.
VICIdial is a trademark of its owner. Trustane is an independent hosting provider.
Sources
- Trustane hopper diagnostic report, read from AST_VDhopper.pl
- Trustane hopper diagnostic specification
- VICIdial Non-Agent API documentation, version function
- Cover photo by Yan Krukau on Pexels
Managed VICIdial hosting


