Skip to content

Current experimental features

> 3.12.0

SC4S_USE_NAME_CACHE=yes supports IPv6.

> 3.0.0

eBPF

eBPF is a feature that leverages Linux kernel infrastructure to evenly distribute the load, especially in cases when there is a huge stream of messages incoming from a single appliance. To use the eBPF feature, you must have a host machine with and OS that supports eBPF. eBPF should be used only in cases when other ways of SC4S tuning fail. See the instruction for configuration details. To learn more visit this blog post.

SC4S Lite

In the new 3.0.0 update, we’ve introduced SC4S Lite. SC4S Lite is designed for those who prefer speed and custom filters over the pre-set ones that come with the standard SC4S. It’s similar to our default version, without the pre-defined filters and complex app_parser topics. More information can be found at dedicated page.

> 2.13.0

  • In env_file, SC4S sets SC4S_USE_NAME_CACHE=yes to enable caching of the last valid host string, replaces nill, null, or IPv4 with the last good value, and stores this information in the hostip.sqlite file.
    • Benefit: More correct host name values in Splunk when source vendor fails to provide valid syslog message.
    • Risk: Potential disk I/O usage and potential reduction in throughput when a high proportion of events are incomplete.
  • To clear hostip.sqlite file, set SC4S_CLEAR_NAME_CACHE=yes flag in env_file. This action will automatically delete the hostip.sqlite file when SC4S restarts.
  • In env_file set SC4S_SOURCE_VMWARE_VSPHERE_GROUPMSG=yes to enable additional post processing and merge multiline vmware events. You should also enable SC4S_USE_NAME_CACHE=yes, to accomodate event that have malformed or missing host names.
  • In env_file set SC4S_USE_VPS_CACHE=yes to enable automatic configuration of vendor_product by source where possible. This feature caches vendor and product fields to determine of the best values for generic Linux events. For example, without this feature the “vendor product by host” app parser must be configured to identify ESX hosts so that ESX SSHD events can be routed using the meta key vmware_vsphere_nix_syslog. With this feature enabled a common event such as an event containing “program=vpxa” will cache this value.
    • Benefit: Less config interaction
    • Risk: Potential disk I/O usage and potential reduction in throughput when a high proportion of events are incomplete.
    • Risk: misidentification due to load balancers and relay sources.
  • SC4S_SOURCE_PROXYCONNECT=yes for TCP and TLS connection expect “PROXY CONNECT” to provide the original client IP in SNAT load balancing.