Why it matters
Busy proxies, NAT gateways and servers under a flood of connections hit the limit first. The symptom is random timeouts, and the kernel log says “nf_conntrack: table full, dropping packet”.
How to check it on Linux
Current entries against the maximum:
cat /proc/sys/net/netfilter/nf_conntrack_count /proc/sys/net/netfilter/nf_conntrack_maxWhat Opservo does about it
The health score flags a conntrack table that is near capacity, before connections start dropping. The Opservo Health Standard
Conntrack table, answered
What is the conntrack table?+
The conntrack table is where the Linux firewall tracks every open network connection. It has a fixed size, and when it is full new connections are dropped.
How do I check the conntrack table on Linux?+
Current entries against the maximum: cat /proc/sys/net/netfilter/nf_conntrack_count /proc/sys/net/netfilter/nf_conntrack_max
How does Opservo handle the conntrack table?+
The health score flags a conntrack table that is near capacity, before connections start dropping.