One of the hallmarks of DNIF Hyperscale SIEM is the ability to collect, parse and enrich logs reliable at high speeds using the lowest hardware footprint in the industry.
Adapters support collection of logs from traditional on-premise models such as UDP/TCP/TLS Syslog as well as custom protocols and cloud APIs.
Recommendations
Our recommended minimum adapter configuration is 16 dedicated vCPUs with 32GB RAM enables reliable collection at upto 15K EPS.
A single adapter can scale vertically and with tuning handle upto 100K EPS.
Adapters have UDP Syslog and TLS PICO Connectors enabled out of the box.
One needs to provision two vCPUs or a single physical core for every additional connector.
Parameter Tuning
Active Time Windows
This parameter lets the system handle late or out of order message processing more efficiently.
Tweaking this is not recommended unless one has multiple PICO components deployed to enable remote collection.
Adapter Pipelines
This parameter lets you scale the number of adapter pipelines from the default 2 upto 8 to more reliably handle larger EPS volumes.
Tweaking this is not recommended unless one observes significant and sustained EPS load.