Memory and CPU bottlenecks were observed with large log traffic. Flush
thread count is increased to 8 and gzip compression is done as a separate
process as recommended in [1], to optimize memory usage. Number of workers
is also increased to 2 to make use of 2 CPUs, to address CPU bottleneck
identified by logs similar to [2] and as recommended by [3]. Unused vars
are removed.
Memory and CPU bottlenecks were observed with large log traffic. Flush thread count is increased to 8 and gzip compression is done as a separate process as recommended in [1], to optimize memory usage. Number of workers is also increased to 2 to make use of 2 CPUs, to address CPU bottleneck identified by logs similar to [2] and as recommended by [3]. Unused vars are removed.