Has any one investigated if influxdb httpd r/w facing scaling issues even when using “batching” with slow down writing to /var/log/messages for the writes.
We start influxdb on centos with std systemd service file and by default it goes to /var/log/messages
Need help to scale the htttp rw and how to move the STD err and STD log to different files in the standard influxdb service files? will that solve the scaling problem . its still a single file for stdout and single file stderr (max)
Appreciate any guidance in this matter.
Do I need to look to the params to scale instead
cache-max-memory-size = 1048576000
cache-snapshot-memory-size = 26214400
I am pretty sure we can scale influxdb but not sure
Attaching the gbd backtrace showing hung in httpd
gdb.txt (28.0 KB)