Unacceptable delay with database 40s

Hei
I am using camunda workflow engine within my spring-boot framework application and I use version 7.15.0 of camunda-engine .
more ever I used Hikari as default spring-boot connection pool and I tried to connect my app to several RDBMS (like Postgres and Oracle)

I have the problem and that is sometimes , operations take so much time to be completed (operations like fetching list of processes and starting process by submitting form and etc …)

I put log on my hibernate level and here is one of them

workflow_1        | 2022-02-05 13:35:47,938 DEBUG org.hibernate.engine.transaction.internal.TransactionImpl [http-nio-9090-exec-3] begin
workflow_1        | 2022-02-05 13:36:16,889 DEBUG com.zaxxer.hikari.pool.HikariPool [SpringBootJPAHikariCP housekeeper] SpringBootJPAHikariCP - Before cleanup stats (total=5, active=1, idle=4, waiting=0)
workflow_1        | 2022-02-05 13:36:16,899 DEBUG com.zaxxer.hikari.pool.HikariPool [SpringBootJPAHikariCP housekeeper] SpringBootJPAHikariCP - After cleanup  stats (total=5, active=1, idle=4, waiting=0)
workflow_1        | 2022-02-05 13:36:17,748 DEBUG com.zaxxer.hikari.pool.HikariPool [SpringBootJPAHikariCP connection adder] SpringBootJPAHikariCP - Added connection oracle.jdbc.driver.T4CConnection@4a738c98
workflow_1        | 2022-02-05 13:36:17,748 DEBUG com.zaxxer.hikari.pool.HikariPool [SpringBootJPAHikariCP connection adder] SpringBootJPAHikariCP - After adding stats (total=6, active=1, idle=5, waiting=0)
workflow_1        | 2022-02-05 13:36:24,279 DEBUG org.hibernate.engine.transaction.internal.TransactionImpl [http-nio-9090-exec-3] committing

as you may noticed , there is about 40 second in performing simple database transaction (fetching list of processes)
and this happens periodically (for example every 40 seconds) .
anybody faced with similar problem ?


results from jmeter has been added

every time delay is happened in application , following log is appeared in console
can’t judge that this is related to the problem
2022-02-12 12:29:12,182 WARN com.zaxxer.hikari.pool.HikariPool [HikariPool-1 housekeeper] HikariPool-1 - Thread starvation or clock leap detected (housekeeper delta=50s783ms660µs445ns).

the problem was not related to camunda at all!
rather I had limited usage of ram in my docker compose
mem_limit: 350m

1 Like

Thanks a lot for posting the solution @Masood_Moghini

1 Like