It takes extremely long (10-20 minutes) to change the view period on the chart.

[microdeth]

Hi, on the historical chart containing the temperature and humidity data when I switch from the last 8 hours to last 3 days, for example, it takes 10-20 minutes. Overall object scheme contains less than 10 data points, only device is connected is our self-designed PLC which have OPC UA onboard.
Thanks,
Dima.

[microdeth]

Server runs at R5 5400U (8GB RAM), client is i7 11700 notebook (16GB Ram). Server = Windows Server 2016, Client = Windows 11

[unocelli]

Hi, sorry for the delay.
we would be very grateful if you would try to investigate in more detail, console error of the frontend and backend, how many tags are you saving? have you tried to re-initialise everything by deleting the current database (_db folder)

[microdeth]

Hi Umberto, Thanks for your response! Overall tags are 8, but only 2 of them are stored:  No errors are seen from Frontend except it asks to wait for the page, the only errors on the backend I've seen so far are:  Now somehow things become much better -- the request to show data for the last 30 days takes only ~7 min (which is not acceptable for the frontend anyhow). I suspect that it can be connected with the overall server performance (it's only a Ryzen 5 5400U CPU with 8 Mb RAM), but I need to confirm that server upgrade may help and this is not some indexing issue. One more note -- when you choose the data range for the chart using the clock icon the dropdown value remains the same, e.g. 'last 8 hours'.  ->  Many thanks for your assistance, Dima. чт, 30 мар. 2023 г. в 01:59, Umberto Nocelli ***@***.***>:

…

Hi, sorry for the delay. we would be very grateful if you would try to investigate in more detail, console error of the frontend and backend, how many tags are you saving? have you tried to re-initialise everything by deleting the current database (_db folder) — Reply to this email directly, view it on GitHub <#746 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6PR3N2NNOH7LCCARIDO6W3W6SICZANCNFSM6AAAAAAWA4TPIY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[unocelli]

Hi, sorry for the delay again. Thanks for your investigation.
I would like to have a look at the archived files. can you share it with me?
have you tried influxdb?

[microdeth]

Hi, thanks for your answer.
You can find the archive files here: https://microdeth.tplinkdns.com/owncloud/s/uIcINRiOBBtkHAA
How can I try 'influxdb'?

Best regards,
Dima.

[unocelli]

Hi, sorry for the delay.
Thanks for the archive files, we will take a look soon.
About influxdb: you have to install it and configure the access value (token and bucket) to insert in FUXA-Settings

[microdeth]

Thanks, will try to! Regards, Dima. сб, 22 апр. 2023 г., 03:10 Umberto Nocelli ***@***.***>:

…

Hi, sorry for the delay. Thanks for the archive files, we will take a look soon. About influxdb: you have to install it and configure the access value (token and bucket) to insert in FUXA-Settings [image: image] <https://user-images.githubusercontent.com/48684550/233734270-c774d081-0402-41ac-9069-3940f7db3684.png> — Reply to this email directly, view it on GitHub <#746 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6PR3N6R53RJWRSZI3NLLQDXCLZVFANCNFSM6AAAAAAWA4TPIY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[unocelli]

is there any news?

[885669801]

有消息吗？

Hello, has your problem been solved? How to solve it?

[andy8601]

I also have similar problems, and the problem of using influxDB is still the same, which can be stored once a second. At the current test stage, there are only about 5 variables, and the data can be refreshed within a few hours. After more than one day, it takes about 15 seconds to refresh the data of the last 8 hours, during which the page is stuck, and the background will not record the updated stored data. It is necessary to wait until the refresh is successful before re-recording the data, influxDB itself has a similar curve chart function, and the refresh speed is very fast, I think there is still something wrong, although I do not understand. If you really can't solve this problem, I suggest that the curve function does not call all curve data by default, and you need to manually check the curve data, I believe it will be greatly improved.

[andy8601]

The problem has always existed. I tried to replace a server with a higher hardware configuration, and some improvements have been made. At present, the data in the last 8 hours of refreshing has been reduced from the original 15 seconds to 6 seconds. In fact, there is only one core in the work of computing, so it caused the delay. I know that node.js has a “Cluster core” module, but I don't understand it very well. @unocelli Hope you solve this problem soon! @unocelli

[unocelli]

Hi, storage on sqlite is file-based, which has several performance disadvantages, it also depends on the split size of the files, the amount of files in the history...I think it would be more acceptable if we split the range into small queries that populate the diagram as in realtime.
Concerning influxdb: currently the queries go through the FUXA server, we should try querying the database directly.

[unocelli]

Hi, The configuration of the DAQ history is decisive. I just added the partial query results for the DAQ data