Hi, everybody. I’m getting close to Grafana as a hobby.
I installed on a raspberry 3B, a sensor ds18b20, nodered, influxdb.
The flow saves the value of the sensor every 5 minutes and the query of the graph of the graph dashboard is updated every 5 seconds.
I have done several tests but every time after about 36 hours the raspberry hangs.
The raspberry is not powerful enough? should I divide the data acquisition with raspberry from saving and displaying with graphana on a PC?
thank you very much
I have a similar system but I save about 20 measurements, some of them every 5 seconds without any problems.
Are you using the Pi to run the browser? I am not doing that, I run the browser on a number of other devices.
After re-booting (assuming that is what you have to do) have a look at /var/log/syslog and see what it shows in the log before the restart. If you can’t understand it then post about 30 lines prior to the restart. [Edit] I should have asked you also to clarify what you mean by hanging. Is the whole Pi hanging so you cannot even ssh into it or ping it, for example.
Thanks for your answer. I hope I have understood correctly what you wrote.
Yes the browser runs on raspberry. I also think it’s too much (ds18b20, influxdb, grafana) for rpi, but on an Italian forum a person using MySQL instead of influx has no slowdowns.
If you run the browser on a PC instead, connected to grafana on the Pi then is it ok?
I certainly didn’t get it right. First the raspberry slows down and then after a few hours still freezes completely.
I have not tried to install graphana on a PC.
I formatted the sd card of the raspberry for other use, given the very long time of approval of my posts on this forum.
Thank you very much
You didn’t need to install grafana on the PC, just run the browser there, connected to grafana on the Pi.
I don’t understand what you mean about approval of your post. You seem to have posted the question 7 days ago and I replied on that day, then you did not post again till a few hours ago.
Thanks for your help but my first two posts were temporaly on hold for several days, I could not do anything on the forum. But not bad.
If I come back to try graphana I will follow your suggestion to use a PC browser.
Have you got the solution to the above problem?
I too had same problem but with twist,
Actually i have influxDB and Graphana installed in the RPi3B+, and run that but as the new user used to view the dashboard, As the number of user increases the Pi got hangged or freezez with time.
It is important to determine whether there is just too much for the pi to cope with (in which case a solution may be to move to a Pi 4) or if there is a specific problem that needs fixing.
The first question is has the pi crashed? Can you still connect via a terminal from another PC? If you can then it has not crashed, run top in a terminal and see what is going on. In fact if you start this while the pi is running ok and leave it running you may be able to see that a particular task is consuming resources.
I hope you are not running a GUI on the pi, that will consume more resources than the rest of the system, so if you are then don’t. You can set that up in raspi-config if I remember correctly, though I always use Raspbian Lite so there is no GUI code at all in the system.
Not necessarily wrong at all, I do exactly the same and a lot more on a Pi 3. You can view the dashboard on any other PC (or mobile) that has access to the network, as you suggested. However if you do want a GUI on the pi then I suggest you should use a Pi 4.
I’m afraid I’m missing some fundamentals. I had understood that they were three well separated parts (node red , db,
Accessing from outside means accessing the DB that resides on the rpi but I see the data with graphana installed on the external device?
Precisely. You can run a browser on any PC on the network and access grafana (or node red) which is running on the pi. At the moment you probably use http://localhost:3000 for grafana when running the browser on the pi, so instead use the ip address of the pi instead of localhost, so maybe something like http://192.168.1.112:3000 if the ip of the pi is 192.168.1.112