Check CONTRIBUTING guideline first and here is the list to help us investigate the problem.
Describe the bug
When upgrading Fluentd 1.6 to the latest 1.9, our soak tests detected that the CPU usage increased by 50% given the same log throughputs. Further investigation revealed the CPU usage increase is triggered by some change between 1.7.4 and 1.8.1.
CPU usage increased by 50%
Memory (RSS) is roughly the same
Log throughputs decreased
The agent can no longer catch up with the same log throughput of 1000 log entries per second:
Expected behavior
CPU usage does not increase dramastically
Your Environment
Fluentd runs inside a Debian 9 based container in GKE with fixed throughput of 1000 log entries per second.
Your Configuration
https://github.com/Stackdriver/kubernetes-configs/blob/b67c14df983107a5eacecf6bdc8371d284aa4d10/logging-agent.yaml#L131
Check CONTRIBUTING guideline first and here is the list to help us investigate the problem.
Describe the bug
When upgrading Fluentd 1.6 to the latest 1.9, our soak tests detected that the CPU usage increased by 50% given the same log throughputs. Further investigation revealed the CPU usage increase is triggered by some change between
1.7.4and1.8.1.CPU usage increased by 50%
Memory (RSS) is roughly the same
Log throughputs decreased
The agent can no longer catch up with the same log throughput of 1000 log entries per second:
Expected behavior
CPU usage does not increase dramastically
Your Environment
Fluentd runs inside a Debian 9 based container in GKE with fixed throughput of 1000 log entries per second.
Your Configuration
https://github.com/Stackdriver/kubernetes-configs/blob/b67c14df983107a5eacecf6bdc8371d284aa4d10/logging-agent.yaml#L131