Cognito brain appliances upload health information to the Vectra cloud using the existing connection to https://api.vectranetworks.com/. This data contains useful information about system health of the brain, sensors, processes, databases, hardware and environment.
Please note: This monitoring is offered at-will by Vectra and does not replace customer-side health monitoring of the platform and service availability. No Service Level Agreements are offered for any monitoring.
The health information is automatically sanitized and checked against known baselines. AI health alerts trigger pessimistically and are often triaged without further action from Vectra or our customers.
Deterministic failures are flagged to an operational monitoring team who validate the findings and investigate the appliance health over remote support VPN (see How do I ensure a Brain has Remote Support enabled?) or a remote support session (see Preparing for a Vectra Support remote session).
Examples of proactive monitoring undertaken by Vectra's operational teams include:
- Hardware monitoring, e.g. disk and power supply health.
- AI performance; indicating unusual traffic patterns or oversubscription.
- Process/service health.
- Metadata upload to Recall; indicating customer metadata bandwidth issue or firewall misconfiguration.
- Disk space availability; indicating oversubscription or software issue causing disk space to be filled unexpectedly.
- vSensor and Stream VM specifications (see Virtual sensor (vSensor) specifications), e.g. incorrect CPU, insuffficient RAM or insufficient disk space.
Examples of proactive monitoring not undertaken by Vectra's operational team include:
- Sensor up/down/available/not available.
- Capture bandwidth, host count and capture performance monitoring.
- Network stability, link stability or network throughput.
The exact monitoring parameters are flexible and change continuously, with feedback from the operational monitoring, security research, engineering and support teams directly driving a fast development cycle to improve the proactive monitoring systems.