Has anyone experience trends only showing on the last portion of the right side?
On the main client it shows correctly, but when opening from other display sizes it looks weird.
Do you have any custom css for that component? Is the trend history shown there compressed or is it clipped?
No custom CSS on the power chart and the trend is clipped.
Here are other examples:
So, it showed clipped when asking for approx. 1 day and then the full trends when asking for several days.
Randomly picking another sensor:
Still clipped, just less.
There isn’t any special configuration, just Ignition 8.3 and its new historian.
Only time I have seen this is when the time on the client is drastically different from the time on the gateway.
1 Like
Just looking at your charts, I would assume you just do not have data over those particular timespans ?
For instance, on the first chart starting at 12am on 03/10, it only shows data at around 9am, so I would assume you do not have data recorded between 12am and 9am.
On your second screenshot, you can see a linear curve (highlighted below)
I’m pretty sure there are in reality only two values recorded for this whole timespan, and Ignition shows a straight line between those two points to fill in the blanks. (Not 100% sure as I can’t really zoom in on the actual chart ahah).
This is something you should check by getting the raw history on this particular tag (using a tag history binding or the system.tag.queryTagHistory function with a returnSize of -1. If there are indeed records “missing”, this is probably why you can see gaps in your chart.
Multiple causes for this :
- Maybe the data is stuck in the store and forward system,
- Maybe your history settings are not ideal for your use case, perhaps you’re using a on change Sample Mode with a historical deadband and no max time between records, which could show as a gap… I’ve seen this happen before for multiple clients because their sensor was giving them a steady value that almost never fluctuates,
- Maybe you lost communication with that particular tag (bad quality, loss of communication with the sensor, etc.).
Without more details on what your history contains and your setup, those are the only things I can think of…
OR maybe I’m completely wrong and it is indeed due to the client time being different from the gateway time…
Looking at the other charts, I would agree with this as well.
This is always an issue with historising slow changing tags if you’ve got it to store values on change. Retrieving the data within a trend window which has no data points within that time, or such as in OP's examples with only data points somewhere within the centre of the window but not the bounds, you will more than likely see blank trend lines to the bounds, as any data points outside the window will not be retrieved…
The fix is unfortunately to set the max time between samples to a value sufficient enough to get you some data points within the shortest window you’ll likely view the data. Eg if you will view an hour, then set it to 20 mins or so to give you at least 2 samples with a 1 hr window.
This will obviously increase your dB storage requirements as well as impact your ability to retrieve only changed data
1 Like