Hi,
I’m using MySQL and Influx persistence in parallel. Im used to Mysql but considering to start using influx. I have created a separate influxdb.persist with a very lean configuration.
Items {
Buiten* : strategy = everyMinute, restoreOnStartup
Binnen* : strategy = everyMinute, restoreOnStartup
Persistence_Every_Change* : strategy = everyChange, everyMinute, restoreOnStartup
}
Every item is persisted once every minute, and a selection of about 50 items are persisted on every change. This is mainly my ESS, Battery, Solar and Grid info. I need those more detailed for analysis. Source is mainly from Victron Venus OS via MQTT. Those update very frequent, like once every few seconds. When I enable the everyChange on the Persistence_Every_Change group I get these warnings:
2025-04-29 09:56:46.872 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:56:49.873 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:56:52.873 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:56:55.873 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:56:58.873 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:57:01.876 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:57:04.872 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
2025-04-29 09:57:07.873 [WARN ] [lient.write.events.BackpressureEvent] - Backpressure[TOO_MUCH_BATCHES] applied, try increase WriteOptions.bufferLimit
When I enable the same in jdbc.persist I do not get those warnings. So i think this back pressure warning has something to do with the Influx persistence binding. All data is stored in docker containers on the same SSD drive, so MySQL and Influx have access to the same hardware thus have the same performance. With iotop I see MySQL eating much more IO then Influx. So I think this backpressure threshold for Influx should be enlarged.
Using Influx v2.
What do you think?
Regards,
Bastiaan