alkismavridis commented on issue #56641:
URL: https://github.com/apache/airflow/issues/56641#issuecomment-3460632269

   @wjddn279 hello! So:
   
   - We run about 120 dags per day (representing about 140 Tasks). All of the 
tasks are Python Operators who wrap some tiny logic around an SSH Operator call.
   - Right, the memory drops represent restarts. As a hacky workaround to the 
memory leak, we introduced a script that monitors the memory in a loop and 
restarts it when: it is above the threshold of 75% AND no tasks are running at 
the moment.
   - Sure! Here it is:
   
   ```
   $ smem -s pss
     PID User     Command                         Swap      USS      PSS      
RSS 
       1 default  /usr/bin/dumb-init -- /entr        0       92      114     
1296 
   19975 default  /bin/sh -c airflow jobs che        0      112      157     
1628 
   19982 default  bash                               0     1700     1760     
3836 
   19990 default  /usr/bin/python3 /usr/bin/s        0    13804    13892    
16412 
    2656 default  airflow worker -- LocalExec        0    26560    42727   
172020 
    2654 default  airflow worker -- LocalExec        0    26676    42847   
172220 
    2645 default  airflow worker -- LocalExec        0    26524    43545   
171808 
    2644 default  airflow worker -- LocalExec        0    26752    43761   
171948 
    2668 default  airflow worker -- LocalExec        0    34172    45770   
170444 
    2676 default  airflow worker -- LocalExec        0    35144    46674   
170588 
    2680 default  airflow worker -- LocalExec        0    36364    47880   
172264 
     856 default  airflow worker -- LocalExec        0    26012    54588   
170760 
      26 default  /usr/python/bin/python3.12         0    14440    55555   
154716 
     857 default  airflow worker -- LocalExec        0    27564    56139   
172308 
     858 default  airflow worker -- LocalExec        0    34768    58567   
169944 
      25 default  airflow serve-logs                 0    20496    61855   
162556 
    9885 default  airflow worker -- LocalExec        0    39616    75244   
178656 
   10202 default  airflow worker -- LocalExec        0    41056    75405   
180696 
   10384 default  airflow worker -- LocalExec        0    41392    75658   
180696 
    9919 default  airflow worker -- LocalExec        0    42316    76601   
178660 
   14485 default  airflow worker -- LocalExec        0    96232   100210   
171112 
   13584 default  airflow worker -- LocalExec        0    96372   100278   
171112 
   12679 default  airflow worker -- LocalExec        0    97780   101180   
171112 
    6250 default  airflow worker -- LocalExec        0    99760   101764   
170236 
    7154 default  airflow worker -- LocalExec        0    99828   101829   
170380 
    7512 default  airflow worker -- LocalExec        0   100668   102598   
170248 
    5359 default  airflow worker -- LocalExec        0   100660   102654   
171148 
       6 default  /usr/python/bin/python3.12         0    96628   102855   
200664 
    8745 default  airflow worker -- LocalExec        0   100804   103257   
172260 
    1752 default  airflow worker -- LocalExec        0   101600   103493   
170828 
    8066 default  airflow worker -- LocalExec        0   102092   104039   
171880 
    3561 default  airflow worker -- LocalExec        0   102140   104078   
171952 
   11108 default  airflow worker -- LocalExec        0   101412   104390   
178860 
   11755 default  airflow worker -- LocalExec        0   101728   104593   
178864 
   10848 default  airflow worker -- LocalExec        0   101804   105288   
180696 
    8980 default  airflow worker -- LocalExec        0   102216   106644   
178720 
    4466 default  airflow worker -- LocalExec        0   105272   107104   
171108 
    8758 default  airflow worker -- LocalExec        0   105852   108276   
177248 
   10128 default  airflow worker -- LocalExec        0   107792   110238   
180660 
   19981 default  /usr/python/bin/python3.12         0   129024   130981   
158464
   ```


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to