Data Usage Leak?

Hi Bob,

My device is remotely located and connected to a 4g modem to do the sending of data. I have been using the emoncms service as a platform to get the logged data out of the device into a web based platform. Running the Emoncms service at "post Interval = 1800sec , Frequency = 6 " ( i.e 1x 30min post with 5min incremental data ) .

The data usage has been exceptionally high at 30 MB per day and i cannot see why as i would presume the "payload to be a few Bytes - ( only 15 values ). The data usage from my ISP comes at a price and i am trying to keep this to a minimum. Ive also tested my data usage with the EmonCMS web server turned off , Updates on Iotawatt set to None , and there still seems to be a minor leak of ( 150KB/hour ).

So , 2 issues in 1 post really. A slow puncture and a very fast one with EmonCMS enabled. Please advise solution Bob.

Many Thanks

A post interval of 1800 sec would be 30 minute data. I don’t know what you mean by “frequency = 6” maybe just post the Emoncms configuration display without the write key so I can see what you are doing.

The only other communication once Emoncms and updates are off would be getting the time once per hour. Those are UDP datagrams and not very big.

Could be that the connection is slow or unreliable and there is a lot of error recovery/retry going on, or it could be that there is some other traffic on the link. You would need to use some kind of external monitoring tool to see what exactly is going on.

The notion of a leak that is somehow flowing out over the 4G link is not realistic. Outbound communications are deliberate and purposeful.

[quote=“overeasy, post:2, topic:1767”]
I don’t know what you mean by “frequency = 6”
[/quote] sorry - i am referring to “bulk send”.

I did some searching on the Emoncms platform and i see there was something picked up in 2017 https://community.openenergymonitor.org/t/emonpi-network-data-consumption/3776 ( not sure if the cause was found but i am going to assume our culprit lies in Emoncms - for the timebeing ).

We had a similar issue with another metering device, where we have over 200 in the field and it was a “handshake not terminating properly” ( can’t recall the terminology ). Appreciating the fact that a few hundred devices start costing fortunes with high data usage…

For now, I think it safest to integrate with the query API as this seems to have the lowest data usage.

I am not clued up enough to understand where the HTTP data usage issues lie but believe MQQT and COap are far more economical.

I will keep subscribed to this post if anyone can shed some light :wink: