You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description:
Apps consumes a lot of memory(RAM) which keeps increasing linearly with time(or events may be).
At start my siddhi app consumes ~320MB and after 5hrs and <120 events the RAM usage goes to >1GB.
After looking into monitoring dashboard it was found that queries with db call(#rdbms:query/cud) and aggregation function (window.timeBatch) consume lot of memory which increases exponentially. Also stream with #rdbms:query for which no event was arrived also consume same memory which keeps increasing.
Database used - Postgres
Events throughput - < 1 event/min.
I suspect the postgres connection is not closed every time a #rdbms:query is requested which leads to memory leak. I have 5 queries in the app where #rdbms:query is used.
Description:
Apps consumes a lot of memory(RAM) which keeps increasing linearly with time(or events may be).
At start my siddhi app consumes ~320MB and after 5hrs and <120 events the RAM usage goes to >1GB.
After looking into monitoring dashboard it was found that queries with db call(#rdbms:query/cud) and aggregation function (window.timeBatch) consume lot of memory which increases exponentially. Also stream with #rdbms:query for which no event was arrived also consume same memory which keeps increasing.
Database used - Postgres
Events throughput - < 1 event/min.
I suspect the postgres connection is not closed every time a #rdbms:query is requested which leads to memory leak. I have 5 queries in the app where #rdbms:query is used.
Affected Product Version:
WSO2 SP 4.4.0.1 (docker)
OS, DB, other environment details and versions:
Tested in multiple environment with different RAM config upto 64GB and Linux and Mac OS.
DB - Postgres
Steps to reproduce:
Connect Postgres to wso2 and create an app which queries postgres using #query:rdbms.
The text was updated successfully, but these errors were encountered: