...
Issue | Notes | Decision | |
---|---|---|---|
1 | Increase memory resources of NCMP (helm chart) | Csaba Kocsis ETH will test and report back to CPS | |
2 | Increase shared_buffer allocation in Postgres config | There is preliminary confirmation that this alone fixes CPS-2156, more testing is in progress | Csaba Kocsis ETH will test and report back to CPS |
3 | NCMP will implement throttling / rate limiting for Rest API (e.g. 503 HTTP response) | Requires determining maximum request rate, e.g. compare previous successful versus failing tests (e.g. 3.4.2 vs 3.4.6) to determine throttling. A poc of rate limiting has been created: 20747: |
|
4 | Rest client (for load tests) will throttle | Depend on outcome of #4 above | |
5 | Lower thread count for Module Sync | This can be done using variable NCMP_MODULES_SYNC_WATCHDOG_ASYNC_EXECUTOR_PARALLELISM_LEVEL (default 10) | Csaba Kocsis ETH will test and report back to CPS |
6 | Review Hazelcast configuration | Hazelcast is configured to have multiple backups which may not be needed given a deployment with only 2 NCMP instances (given 2 instances, only 1 backup would be needed). Testing has shown this reduces heap usage by around 100MB during 20K CM handle registration. | Daniel Hanrahanhas provided a patch to reduce memory consumption: https://gerrit.onap.org/r/c/cps/+/137517Daniel Hanrahan will report findings |
Background
CPS and NCMP have much higher memory consumption than required. Regarding NCMP specifically, it has some in-memory data structures that grow linearly with the number of CM-handles.
...