Troubleshoot and resolve VICIdial issues

VICI DIAL ISSUES RESOLVE

# VICIdial — Current Issues Report

**System:** VICIdial + Asterisk/PJSIP
**Agent tested:** `101`
**Primary objective:** Stable agent connection + reliable auto-dialing

### 1. Agent is becoming `LAGGED` — Critical

Agent `101` is being shown as:

```text
101 PAUSED LAGGED
```

The `LAGGED` state occurs because VICIdial's backend determines that the agent's state/heartbeat has become stale.

---

### 2. `last_update_time` stops updating — Critical

We observed that the agent's `last_update_time` becomes significantly older than the current time.

This indicates that the agent's normal heartbeat/update mechanism is **not continuously reaching or being processed by the server**.

This is currently one of the strongest pieces of evidence in the investigation.

---

### 3. Browser-side JavaScript/AJAX heartbeat failure — Critical

We traced the failure chain to the agent interface:

```text
Browser JavaScript
↓
AJAX request
↓
xmlhttprequestcheckconf
↓
vicidial.php
↓
JavaScript/PHP response problem
↓
heartbeat/refresh processing fails
↓
last_update_time freezes
↓
ADMIN_keepalive_ALL.pl
↓
LAGGED
```

The browser-side code was encountering an:

```text
undefined.split()
```

type of error.

This is particularly important because it gives us a **specific technical path to investigate**, rather than treating `LAGGED` as the root cause.

---

### 4. `vicidial.php` around line ~2130 requires investigation — High

We previously narrowed the problematic processing to approximately:

```text
vicidial.php
around line 2130
```

The issue appears associated with a value that is expected to contain a string but is instead arriving as `undefined`/missing before `.split()` is called.

**Root cause has NOT yet been conclusively identified.**

So we should **not modify this code blindly**.

---

### 5. `ADMIN_keepalive_ALL.pl` is detecting the stale agent — Confirmed behavior

We found the relevant stale-agent handling in:

```text
/usr/share/astguiclient/ADMIN_keepalive_ALL.pl
```

This process is detecting agents whose update timestamps are too old and subsequently treating them as stale.

This is **probably not the root cause**.

It is more likely the component that is **correctly reacting to the earlier heartbeat failure**.

So:

> `ADMIN_keepalive_ALL.pl` → symptom/enforcement
> Heartbeat failure → likely root cause

---

### 6. Auto-dialing is therefore unreliable — Critical

Because the agent is becoming `LAGGED`, the predictive/auto-dialing process cannot reliably maintain the agent as available.

We have confirmed that the system has leads in the hopper, so the problem is **not simply "there are no leads."**

The current priority is therefore to stabilize the agent state before changing dialing parameters.

---

### 7. Root cause is still NOT conclusively identified — Important

At this stage, we should **not claim** that the problem is definitely:

* Asterisk
* PJSIP
* SIP provider
* MySQL
* network
* firewall
* `ADMIN_keepalive_ALL.pl`
* dialer settings

We have stronger evidence pointing toward the **VICIdial web-interface heartbeat/AJAX/JavaScript processing path**, but we still need to identify exactly **why the expected value is missing/undefined**.

---

# Current Status

| Component | Status |
| ------------------------ | --------------------------------- |
| VICIdial installation | Working |
| Asterisk | Running |
| PJSIP | Working |
| Agent 101 | Can operate but becomes LAGGED |
| Leads/hopper | Leads available |
| Agent heartbeat | Failing/interrupted |
| `last_update_time` | Freezes/stales |
| `ADMIN_keepalive_ALL.pl` | Detecting stale state |
| Auto-dialing | Unreliable |
| Exact root cause | Still investigating |