Hey All,
I built out my Node Red system 6-8 months ago and have been happily using it ever since. I am doing some graphing but I balanced datapoints to performance months ago.
On a recent rainy day I discovered that you can collapse Groups on the dashboard. So I went through and did that to the 10ish groups I have. This seems to have opened up a memory leak in Chrome causing slow performance and eventually a message saying I've used up available memory. So I went back and turned off the collapse feature and performance returned...sort of.
I access this dashboard from two Windows 11 machines and an iphone. Things run well on the iphone and one of the Windows machines but not the other. They're both using Chrome, I cleared the cache on the problematic machine but it still hangs and sucks up the memory. I poked around McAfee software but didn't find anything blocking.
How can I get this machine to render the dashboard again?
I checked out the console on the good machine which of course showed issues with the graphs/charts. So I disabled the graphs and the dashboard works speedy everywhere.
Several months ago after I discovered the limitations of charts in Node-red I looked for an alternative method. I explored a different system with a database and JQuery or something, I can't seem to remember the name. I had trouble getting it configured, got frustrated and decided to bite the bullet and stick with the node-red charts.
I'm still interested in a database based charting system for the dashboard. Anyone have suggestions?
The node red charts are ok provided you keep the total number of data points in all the charts to a sensible number. That means keeping down the total time range of each chart and keeping down the the data update rate for each line.
The sort of thing to avoid is a chart with a time range of 24 hours, containing 5 lines, each updated every second. That means the after it has filled up the chart will contain 5*24*3600 stored data points, which is 432000. Every second the browser will have to shuffle those 432000 points down the chart, which is a lot of work to do. If you have a lot of charts then that multiplies the problem up.
The hardest part is getting your head around InfluxDB and configuring some sensible data. I know I posted some guidance on that quite a long time ago. You do also need to keep your InfluxDB databases to sensible sizes as well, especially on a Pi.
But InfluxDB has some built in features for trimming DB's but also for aggregating data. For example, I keep my once-per-minute sensor data for a week but also aggregate to hourly data as I go along and keep the hourly data (with min/max/avg data points) for a few years. InfluxDB has features for automating all of that.
