Neptune Apex plugin - unable to parse "none -0400"

[boozer2]

Relevant telegraf.conf:

# Neptune Apex data collector
[[inputs.neptune_apex]]
   ## The Neptune Apex plugin reads the publicly available status.xml data from a local Apex.
   ## Measurements will be logged under "apex".

   ## The base URL of the local Apex(es). If you specify more than one server, they will
   ## be differentiated by the "source" tag.
   servers = [
     "http://192.168.111.111",
   ]

   ## The response_timeout specifies how long to wait for a reply from the Apex.
   #response_timeout = "5s"

System info:

Telegraf 1.10.4
Ubuntu 18.04 Server

Steps to reproduce:

I haven't gotten this to work at all because of this parsing error. This appears to be because the plugin is expecting a date but the Apex is returning "none" for power restored time.

Expected behavior:

No errors, and data logged to InfluxDB

Actual behavior:

Error: telegraf[9919]: 2019-05-21T00:33:50Z E! [inputs.neptune_apex]: Error in plugin: unable to parse "none -0400" (parsing time "none -0400": month out of range)

And nothing is added to my database from this input plugin.

Additional info:

Beginning of my Apex status.xml:

<status software="4.53_AB17" hardware="1.0">
    <hostname>FishTank</hostname>
    <serial>AC4:12345</serial>
    <timezone>-4.00</timezone>
    <date>05/20/2019 20:24:36</date>
    <power>
        <failed>05/06/2019 17:43:41</failed>
        <restored>none</restored>
    </power>

[danielnelson]

It seems like we should either ignore errors parsing the failed/restored time and possible log them as warnings (could be annoying if restored is always none). @MaxRenaud What do you think?

[MaxRenaud]

Yes, this works. I'm surprised that this field can ever be "none" considering the device has power and serving an xml file.
I'll submit a diff this weekend that ignores errors in both "failed" and "restored" and omits these fields. I'm a bit hesitant to log them as warnings since it seems like a persistent error but I'm OK to do it if it fits Telegraf's general logging practices from external sensors

boozer2,
I'm sorry this is happening on your new installation. As a temporary fix until we get a new release done, you can try to power cycle your apex from the wall and hopefully this field will be populated in the status.xml

[danielnelson]

If the error is persistent, and can't be cleared up by the user, then we shouldn't log.

[boozer2]

I power-cycled, and I see the "failed" time has updated, but the restored time still says "none". That is kind of strange...

[MaxRenaud]

@danielnelson ACK for the logging
@boozer2 This seems like an Apex bug and we'll work around it by not logging these invalid fields. I imagine any automation will have do something like "if current_failed > previous_failed" or something like that so the lack of data should be OK.