SCADA database grew to 180 GB in a year: which log settings should I check?

[SignalFlow]

Our Omnivex SCADA project has around 1,500 tags. The server's disk is almost full and our IT department found that the database has grown to about 180 GB in one year. Reports with longer time ranges have also become slow.
When the project was created, I think most tags were simply set to periodic logging with the same short interval. I don't want to lose important history, but clearly we log far more than we need. How should I review the log settings, and what values make sense for different types of signals?

[Technical Support]

Your suspicion is very likely correct. Database size grows with the number of logged tags multiplied by how often each one is recorded.
Settings that control logging
- Log Status: Periodic records the tag; No Log does not.
- Log Time: the logging period in milliseconds. 1,500 tags at 1000 ms produce many times more rows than the same tags at 60,000 ms.
- Filter Active / FilterMinValue / FilterMaxValue / FilterInOut: filter which values are recorded, useful against invalid readings.
- For alarms, Log Active in the Alarm Editor decides whether alarm states are written to the database.
How to review
1. Export the tag list (File > Import/Export Tags > Export) to get an overview in Excel with Log Status and Log Time per tag.
2. Group the tags:
- Not needed in history (internal helper values, duplicate signals, button states): set to No Log.
- Slow process values (tank levels, temperatures, energy counters): long periods, for example 30–60 s or more.
- Fast or critical values you analyze in detail: keep shorter periods, but only for these.
3. Change the settings in the Tag Editor, starting with the tags that have the shortest Log Time.
Important
- Changing the settings reduces growth from now on; it does not shrink existing data. Cleaning up or archiving old records is a database task. Coordinate it with IT and take a backup first.
- Before changing anything, check which tags are used in saved reports and templates, so reports don't lose data.
Documentation: https://greensoft.ai/documentation/o...ion-protocols/