Tag history binding slows system

Ignition
1

Hi,

I’m working on some trending views and currently facing some performance issues.

I have one view where I show 8 trends with each 2 pens

When I open the view with the trends, the gateway cpu starts spiking to 70% where it is at 6% when viewing a screen that does not have trends.

The query throughput of the core historian (historian.IgnitionCoreHistorian.query.throughput.actual) average at 400 queries per second (0 when not viewing trends)

This seems excessive to me for 16 pens x 100 points

Anyone a idea what I’m doing wrong?

Using the config:

image
image874×536 29 KB

custom.output is : [

{

"aggregate": "SimpleAverage",

"alias": "PV DCI Speed",

"path": "\[default\]204/501/011/Act_Speed"

},

{

"aggregate": "SimpleAverage",

"alias": "SP DCI Speed",

"path": "\[default\]204/501/011/Act_Speed"  }

]

I don’t think it’s a logging problem but here is the logging config

image
image436×297 11.4 KB

2

Don't use realtime mode for many pens, unless you set the chart to update very slowly.

(Ignition queries the entire realtime span for every update. It's caching system can't really handle this.)

1 Like
3

does the system.historian.queryRawPoints also queries the entire realtimespan? the disadvantage of the historical tag history is that it does not update every so often

4

Not sure. That function is new for v8.3 and I haven't dug too deeply into it yet.

5

This would be lovely, and bring it on par with other scada trending charts.