Hi all. I have an issue with the built in Dashboard UI charts frequently re-setting throughout the day, limiting the ability to see trends. I am using them for temp and humidity monitoring, however con only ever get 6 hrs historical data at a maximum.
Note:
The charts are set to display 24 hours of data.
The node-red server and service both have uptimes greater than 24 hours.
There are two likely possibilities that I can think of. Firstly it is possible that node-red (or the machine) is crashing and restarting for some reason, possibly because you have too much data on the chart. You can test this by starting node-red in a terminal and leaving the terminal open and seeing what happens when the chart gets cleared.
Secondly it may be that you are sending it an empty array (or something similar) instead of just values, which will have the effect of clearing the chart. Do you have anything in your flow that could do that?
In terms of the server or service crashing, I have monitors configured to check uptime of the service and it does not reflect any restart. I will, however, open a terminal as suggested to see if any activity aligns with the reset.
There is nothing in the flow that would send an empty payload, however I may log the output to a file to ensure this is the case.
You can test the limits of the browser by creating about 10 charts, update them every 0.1 seconds and and watch couple of minutes until the browser crashes. Keep some timer running alongside. Then do the math to figure out if you are near to that limit with your long running charts.
It is not most accurate test but gives some kind of estimates.
There is actually more conditions like data amount over the socket in different situations and so on. But it all ends with amount of the data points to chart to be calculated and rendered. And that is mostly the case.