Hi,
Lets say I have an OPC UA server (kepware), which has 30.000 tags defined, and I want to push all these (ideally) into InfluxDB, using the OPC UA Listener plugin in Telegraf. How would I structure this the best way? Would a single Telegraf process handle this? Is it considered good practise to have such a big configuration file? Should this be split into multiple Telegraf processes and files? Would this be better solved via the Kepware IOT gateway (using rest client)? I want to be able to make a solid solution, which scales and would be robust. I dont have a problem maintaining big config files, as htese can be automated, but I need to figure out what is the best practise and how I would go ahead and configure this. Thanks
Do you know how many metrics per second you would expect to send to Telegraf? Will you also be doing any processing of the data with telegraf?
Is it considered good practise to have such a big configuration file?
There should be no issues with a large config file for telegraf itself. It is more a management concern, where you would want to use some sort of configuration or templating to manage the config file.
Should this be split into multiple Telegraf processes and files?
Multiple files makes no different to telegraf itself. All files are combined into one. Splitting this up into multiple processes would be dependent on my question about messages per second. You would want to see how we react to your data first.
I will not process any of the values, only send them directly to InfluxDB. I`m not sure how many tags will be sent each second, frankly. All I know is that the usecase is around 20-30.000 tags (probably alot less usefull ones), and I want to see if it is required to split this into multiple Telegraf processes, or if a single one, can handle this, configured with multiple OPC UA inputs definitions. Would a single Telegraf process be able to hanle this? either via the OPC UA cyclic read plugin, as well as the OPC UA Listener plugin?
Just out of curiosity, how many tags are considered way to many, for a single Telegraf process?
Without knowing how many messages per second or some other values, I am not sure. I do not have hard data to say one way or another. The only thing I have seen, which is from the OPC UA library we use was a discussion post asking about reading 60k tags:
