BrainFrame Forum

Why do previously triggered Alerts get deleted from time to time?

The actual measurement found that the information in the alarm (alert table) generated by the system will be deleted from time to time. It is not sure what the deletion mechanism is. Generally, the alarm in the morning will be deleted in the afternoon.

From the log point of view, [36mcore_1 |^[[0m INFO:root:Pruned 162477 zone statuses, 5840 detections, 3 alerts, 3 data stores in 38.32919669151306 seconds, it feels that it was deleted when the judgment was made based on certain rules
logs.txt (385.0 KB)

@alex.thiel


系统触发的告警信息会被不定时删除

实测发现,系统产生的告警(alert 表)中信息,会被不定期的删除,不确定删除机制是什么. 一般早上的告警下午就会被删除

从日志角度看,[36mcore_1 |^[[0m INFO:root:Pruned 162477 zone statuses, 5840 detections, 3 alerts, 3 data stores in 38.32919669151306 seconds ,感觉是基于某些规则做判定时删除的

1 Like

Hi there,

Thank you for posting on our forum! THe mechanism you are talking about is called “Journal Pruning”.

It can be configured by creating a file called “.env” in the same directory as your “docker-compose.yml”.

cd $(brainframe info install_path)
nano .env

Then adjust the values described in this documentation:

Basically, JOURNAL_MAX_AGE=60d0h0m means that BrainFrame will delete all data over 60 days old.

Then, we have PRUNING_AGE and PRUNING_FRACTION.
If PRUNING_AGE=0d2h0m, and PRUNING_FRACTION=0.05, that means that BrainFrame will delete 5% of data older than 2 hours, each hour. This is a way of reducing the size of the database.

If you want to disable pruning and keep 7 days worth of data, you can do the following:

PRUNING_FRACTION=0
JOURNAL_MAX_AGE=7d0h0m

PRUNING_FRACTION=0 means that 0% of data gets deleted each hour.

Thanks!

I tried to make a new capsule "classier person attributes ”, while running it on brainframe, an error occurs like “Cannot create ShapeOf layer 452/input_rank/shape_of id:345”

how to fix it? I use the pre-trained model person-attributes-recognition-crossroad-0230 in master branch

Sorry to see you had a problem with the “classifier person attributes”. I cannot seem to reproduce the issue. Try using the packaged capsule I have included here and see if you still get the same issue. classifier_person_attributes_openvino.cap (1.6 MB)

If possible could you post your whole error stacktrace? And maybe provide a zipped directory of the capsule you were trying to use.