[Ovmsdev] Queue overflows

[Michael Balzer]

This normally happens if the Wifi connection isn't fast / stable enough 
for the WebSocket stream(s), or when the web server doesn't get enough 
CPU share.

It also depends on the frequency and volume of changes in the metrics & 
logs, which need to be transported to the web client(s).

I had taken some measures to cut down the logging for these issues, but 
I think that needs to be optimized further -- especially as logs now 
also are transported to the web clients. If for example you've got two 
web sessions open and the first runs into the overflow, the second will 
be flooded with the warning messages and probably run into overflow as 
well. I'll put that on my list.

Steve, you can try raising the queue size, but I think you didn't find 
the right queue. The default size is 50, the queue is created in the 
WebSocketHandler constructor (m_jobqueue).

If that helps we can raise it generally or add a config for this.

Regards,
Michael


Am 11.01.21 um 11:15 schrieb Chris van der Meijden:
> Hi Steve
>
> I know that one too.
>
> Happens to me always, when I get CAN Errors when I wake up the CAN 3 
> Bus on the VW e-Up.
>
> Greetinx from Bavaria
>
> Chris
>
>
>
> Am Montag, den 11.01.2021, 11:44 +0200 schrieb Steve Davies:
>> Hi,
>>
>> I generally debug by watching the shell via the web interface with 
>> logging from my vehicle module at debug or verbose.
>>
>> I seem to be hitting throughput limitations for the websocket pushing 
>> the output to my client, though, and see many of these:
>>
>>     log.20210110-144200:2021-01-10 14:17:39 SAST W (1630612)
>>     websocket: WebSocketHandler[0x3f8531a4]: job queue overflow
>>     resolved, 133 drops
>>
>> This especially happened when I went to a charger for a test and was 
>> connected to the OVMS system on 192.168.4.1 on the local Wifi AP that 
>> the OVMS creates.
>>
>> Is this to be expected?  Is it more efficient to "log monitor" via 
>> ssh, or do I really need a serial connection over USB?
>>
>> I do see that the websocket overflows seem to occur when any client 
>> is connected whether you are looking at the shell or not.  In my case 
>> I'm logging quite a lot and I guess I exceed the 100 deep queue.
>>
>> I see that I can make the queue bigger for the websocket - would it 
>> be a big problem to increase the 100 to 200 by default?
>>
>> Thanks,
>> Steve
>>
>> _______________________________________________
>> OvmsDev mailing list
>> OvmsDev at lists.openvehicles.com  <mailto:OvmsDev at lists.openvehicles.com>
>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev  <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
>
> _______________________________________________
> OvmsDev mailing list
> OvmsDev at lists.openvehicles.com
> http://lists.openvehicles.com/mailman/listinfo/ovmsdev

-- 
Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
Fon 02333 / 833 5735 * Handy 0176 / 206 989 26

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20210111/534705f8/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature
Type: application/pgp-signature
Size: 203 bytes
Desc: OpenPGP digital signature
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20210111/534705f8/attachment-0001.sig>