why is it that the flukso graphs are anywhere from 10-15 minutes behind real time unless you open minute view (but cant on my phone as i cant click load insecure script)?
that then makes the pvoutput 15-20 minutes behind
when i was using a CC ENVIR the pvoutput graph was done basically straight after the 5 min Poll Interval ie as close to real time as possible
Do you mean the graph time is behind or if the measurements shown are shifted?
Yes the graph doesn't get updated until 10-15m after the recordings
See http://pvoutput.org/help.html#autoupload-flukso - and report to pvoutput.org as this is no flukso.net issue.
@gebhardm
its the flukso graph that is 10-15 mins behind real time.
that then makes pvoutput 15+ mins behind
https://dl.dropboxusercontent.com/u/58260317/Flukso/Time%20delay.png
Ooops... I didn't see this was about pvoutput. That's now what I think of 'flukso graphs' ;)
i was talking about flukso graph im assuming flukso only updates to the dashboard every 10 or 15 mins?
hence why the pvoutput is 15- 20 mins behind
OK - "dash" it is called in the menu of this site, that would have been distinct - there are by now so many "flukso graphs"...
Ref to https://www.flukso.net/content/api-flukso-apifluksonet and some guess what happens:
* The FLM should push its data to the FLM server every 300 sec (see config - this is to not overload the "free" service provided by icarus75...)
* The internet is not a "real time" infrastructure, so you have to expect delays
* The FLM server is in Belgium, so some distance, thus "hops" away from your location
* pvoutput pulls every 5 min
* so, with some latency assumptions you are easily at the 5 min push to Belgium plus 5 min pull back to the other side of the globe - thus, easily 10 min difference...
And hey, I am also just a Fluksonian ...
I understand.
I was just saying that flukso can be anywhere from 10-15minutes behind just wondered if it was just me, a bug or everyone
well as I wrote, the flukso chart on flukso.net (aka the Dash) is never more than 5 minutes old, in fact you see the time grid go all the way to the current minute, and 0 to 5 minutes of graph lines missing.
I do not think that distance or hops have anything to do with the problem (if they do, you or your internet connection have bigger problems), access times to servers across the planet are in the 100ms range these days.
100ms is just the ping time, not the time until you can expect an ACK (or any fulfilled complex message transmit)... (otherwise I would not understand why our cloud people cry when they have to access European data centers from South America) - as I wrote, just a guess; a traceroute may show more...
As a last response (as we drift off-topic), let me say that one transaction of pushing values to the server should not differ much from that 100ms.
Now logging in and working with servers that have a delay of 100ms for every transaction, I agree it will be slow, and that I can understand... ;)