Last night I changed my sitemap so I had both a webview and static render of the same data, with IP address specified. Trying now with myopenhab and the android app, myopenhab shows only the static render, the android app shows neither, so the IP specification and statis render still doesn’t get you any closer in the app.
Hi Kevin, I had mine set up as per the influxdb+grafana persistence tutorial, which only specifies the first two changes. I’ve just gone in and changed [auth.basic], which was commented out and set to enabled=true.
I thought that it worked, getting both static and webview showing up on my android tablet, but then I realised it was on the local network. Tried my android phone over the internet and still no go unfortunately, and via myopehab.org on a browser shows the static version.
I will re-check my setup and see if I can figure out why I’m getting a different result based on the the static image url.
Update: Just re-tested and Android is not working when on mobile data as I thought it was, works fine on wifi, I’m getting the following errors in the opehab log but not sure what they relate to at this time, too late to take a look tonight.
Another update: Again tested from Android and the static Image graphs do show thye are just very slow this is via data with wifi turned off so hitting remote url.
The same messages still appear when accessing via Android app but everything works at least for me.
Hi Kevin, I too just figured out that it’s taking a very long time to load but does get there eventually; accidentally found this when I left it open and went off to do something.
That said I just tried it multiple times, had a few timeouts then the odd quick load, so who knows
Thats good news, in a way, as we are both seeing the same thing.
One thing to watch for is the number of open connections to your Grafana service can get get out of hand and flood the service if you refresh or try to open links on multiple devices, my Grafana/InfluxDB service is on another Pi3 and I have seen that machine grind to a halt because OH had over 65 established connections on port 3000 this caused the Pi to become un-responsive and only a hard reboot would fix the issue, no access via ssh when this happens, I now capture the number of connections on port 3000 and alert when they start to get to high.
Ive tried adding tcp parmeters to sysctl.conf and also disabled ipv6 connections but nothing thus far has stopped this from flooding the service.
Ok after having to yet again force a power off the grafana Pi3 because OH generated 60plus sessions for a few graphs I was trying to display I have now taken a different approach to displaying the graphs in OH.
I have created a comma delimited file which contains , I then have a small script which essentially reads the previous file and performs a wget of each graph and saves them in a gimages folder under /etc/openh2/html.
The wget script is cron’d to run ever 5 minutes, it takes around 200 seconds to run and generate 18 images.
I then changed all the static Image urls to point to the pre-pulled images in the gimage folder.
The graphs now load really quickly in OH web browser and work fine for the IOS and Android apps when accessing the local url.
Remote URL on IOS works, sometimes really quick and sometimes slow.
Remote URL on Android is still very much hit and miss but more miss than hit.
I still get the following errors for both IOS and Android while using remote URL.
I have gone for this solution because of what I’ve explained above, using the image renderer does work however on several occasions just looking at these graphs had resulted in lots of connections 60+ being established to my Graphana/InfluxDB pi ending up with no graphs and nothing getting captured in the dB from OH, I then have no choice but to power off the pi3 as I can’t even get ssh access.
If you have an idea how to stop this behaviour then please share otherwise I’m stumped and my complicated solution will just have to do for me.