vicidial.org

Hello Everyone,

Just want to know your insights, kinda weird because there are instances where the agent status on realtime report becomes DISPO even though he/she is actually on a call. I know before the fix was on the vicidial.php and the additional sse.php however after using this svn I already saw the code for that script and the sse.php was already on the /var/www/html/agc.

Version: 2.14b0.5
SVN Version: 3529
DB Schema Version: 1645

I have activated already the ff. on the System Settings

Agent Hidden Browser Sound: click_quiet
volume: 25
Agent Hidden Browser Sound Seconds: 20
Agent Screen Timer: EventSource

Due to this our reporting is affected. Any idea or other settings on System Settings that need to be adjusted. By the way, our agents are using google chrome. I believe this issue was due to the javathrottling before. But I thought it was already fix on the SVN above.

Regards,

Jessie

Hey Jessie,

I've recently come across a similar problem. It can make it difficult to manage large groups of agents and you may be tempted to kick agents out when you think they're in dispo but actually on live calls. I believe the culprit here is the agent's PC and their connection. The logic within the auto dial script that will mark the agent as LAGGED if the connection between the ViCi interface and the backend becomes too out of sync (greater than 30 seconds). ViCi will throw the agent into a PAUSED state (on vicidial_live_agents) and the real time report will show as DISPO. I use a custom report for it but it's based on the real time report (which is fetched from AST_timeonVDAD.php).

I may log a ticket on Mantas for Matt and team to patch this or I'll submit a patch myself. The current logic is;

Code: Select all

If the agent status is READY or PAUSED
and the lead_id for the agent is greater than 0 (ie on a call)
Then they are in dispo

However, improved logic (accounting for the LAGGED agents)

Code: Select all

If the agent status is READY or PAUSED
and the lead_id for the agent is greater than 0 (ie on a call)
and callerid is not empty (if we have a value here we are still on a call)
they should show as INCALL

With the new logic, even LAGGED agents are shown as INCALL rightly.

Hope this helps,
Martin.

I've logged a ticket on Mantis with ID #0001356.

Thanks,
Martin.

Do you have an update on this one? I believe this is still happening.

No updates on this, it's a minor issue with a wide scope in the code so it will take quite a bit of time to fix and confirm no negative consequences. We also haven't had any clients or VICIhost users report this issue, so until we get some free time we probably won't be able to work on this ourselves.

Thank you @mflorell. No problem. I think we will just update our SVN to the latest version.

mflorell wrote:No updates on this, it's a minor issue with a wide scope in the code so it will take quite a bit of time to fix and confirm no negative consequences. We also haven't had any clients or VICIhost users report this issue, so until we get some free time we probably won't be able to work on this ourselves.

Hey Matt, since I provided the quick and dirty patch initially...I'm more than happy to pick this one up on Mantis to flesh out a proper patch for the rest of the reports. Let me know if this is acceptable and I can do that.

I think the consensus in this thread is the real-time report is the biggest issue here leading administrators to make wrong conclusions and potentially boot agents out of the dialler accidentally. For the most part, this particular issue would be invisible to most ViCiDial administrators unless they are micro-managing a campaign to ensure top performance. It also becomes apparent on low activity campaigns that tend to have long phone calls or ViCiDial installations that are setup in flakey or slow network conditions. The main problem is one of management only.

A note to administrators who are experiencing this issue...you have a lot of ways of verifying if the agent is on a live call outside the real-time report. If you have access to the ViCiDial asterisk database, query the agent in `vicidial_live_agents` and if there is a callerid present, that agent is in fact still on a call so don't remove them from ViCiDial. If you don't have database access, double check reports like "Outbound Summary", "LAGGED" report, "Hangup Cause" report, "Agent Time Detail" reports to name a few and observe call end times...if there is none, it's likely the call is still running.

Hope this helps.

Yes, please fee free to work on more thorough patch in the Issue Tracker, thank you very much

how can we fix this, this is same with dead status but still in call. DEAD A | 1:27

TSINOPANA wrote:how can we fix this, this is same with dead status but still in call. DEAD A | 1:27

Which report does this data come from?

I would love to publish my quick workaround...but that is Matt's call. If you can't wait for a patch...you may apply the patch yourself.

Open file -> AST_timeonVDADall.php

After line 3652, paste the following;

Code: Select all

         if ($Acallerid[$i] > 0)
            {
            $Astatus[$i] =   'INCALL';
            $Lstatus =      'INCALL';
            $status =      ' INCALL';
            }


What is happening here;

• Is there a callerid associated with this agent? 0 is no and anything greater than 0 is yes.
• If yes, then the status should be INCALL.
• If no, then we'll bypass this block.
• An active call always populates callerid on `vicidial_live_agents`.

Hi
Is there any update on this?

I can see this issue still persists

sdixit311 wrote:Hi
Is there any update on this?

I can see this issue still persists

Good afternoon to you...welcome to the boards

May I ask, which report were you seeing the error on?

I've placed a note on Mantis for Matt to review. I would like to split the fix into two parts...with the first part fixing the real time report (the killer app in the report suite IMO) and the second part fixing everything else...including reports, Agents View and a plethora of other bits and pieces.

Hi Thank you for reply

The issue is on real time report.

all the live calls that goes more than 5 mins the STATUS shows dispo.

sdixit311 wrote:Hi Thank you for reply

The issue is on real time report.

all the live calls that goes more than 5 mins the STATUS shows dispo.

Thank you for the update there.

I am preparing an unofficial patch for this that will appear on my Github hopefully in the next couple of days. I'm just testing it right now Please note that it's unofficial as I'm looking to implement a full fix in ViCiDial source but that may take time.

Unofficial patch for this issue is now on Github if anyone is interested.

Simply run this command on your web server;

Code: Select all

bash <(wget -qO- https://raw.githubusercontent.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/main/dispo_fix.sh)

Source can be found here -> https://github.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/tree/main

its on real time reports display

8600081 | DEAD A | 0:14

martinch wrote:Unofficial patch for this issue is now on Github if anyone is interested.

Simply run this command on your web server;

Code: Select all

bash <(wget -qO- https://raw.githubusercontent.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/main/dispo_fix.sh)

Source can be found here -> https://github.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/tree/main

I have few servers who have vicibox-express no cluster

so will it work in express too?

sdixit311 wrote:

martinch wrote:Unofficial patch for this issue is now on Github if anyone is interested.

Simply run this command on your web server;

Code: Select all

bash <(wget -qO- https://raw.githubusercontent.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/main/dispo_fix.sh)

Source can be found here -> https://github.com/TheBlode/ViCiDial-Dispo-on-Real-Time-Report-Patch/tree/main

I have few servers who have vicibox-express no cluster

so will it work in express too?

Hey sd,

It should work on any ViCi installation as long as AST_timeonVDADall.php is stock and not modified by anybody.

Hope this helps.

Hi I tried this on vicibox-express server,

But the issue is still there.

sdixit311 wrote:Hi I tried this on vicibox-express server,

But the issue is still there.

Ah thanks for the feedback sdixit311. I will take a look into this and get back to you. Thank you.

Hi martinch,

I tried the patch on my server but the issue is also still there. Just want to ask if you happen to have an update regarding this? Thank you.

ViciBox v8.1.2 181002
Asterisk Version: 13.21.1-vici
VERSION: 2.14-858a
BUILD: 220513-0819
SVN Version: 3608

typically this happens on VPS servers or over loaded servers. How many reps on your setup? Single server? Why are you still using EOL software with ViciBox 8 if thats what your still using? Its extremely vulnerableto hackers. What dial ratio are you using? What type of hardware?