This week’s GitHub Project of the Week is Vector, a newly open-sourced tool from Netflix for host-level performance monitoring.

Vector is a framework for exposing handpicked high-resolution system and application metrics to every engineer’s browser. It implements the open-source Performance Co-Pilot (PCP) framework with a new UI and configurable cross-metric dashboards, allowing users to visualize and analyze system and application-level metrics in nearly real time.

“Having the right metrics available on demand and at a high resolution is key to understanding how a system behaves and correctly troubleshooting performance issues,” Netflix engineers Martin Spier, Amer Ather and Brendan Gregg wrote in a blog post.

Vector is built with Angular, with the D3.js data-driven document library for JavaScript integrated for charting capabilities. The Web application has a “default” launch dashboard to control a given system, and uses a data model prototype called WidgetDataModel to enhance function customization for reading and analyzing CPU, memory, disk and network metrics.

Netflix’s plans for Vector going forward include adding more widgets and user-defined dashboards, metric snapshots, CPU flame graphs, disk latency heat maps, and support for Cassandra. Vector depends on PCP 3.10 or higher to run.

“Observability is key to understanding how an application behaves under certain conditions and is paramount to successfully troubleshoot any performance issue,” the Netflix engineers stated. “Vector allows us to closely monitor hosts in near real time and easily correlate metrics, making them accessible to every engineer, simplifying the process of troubleshooting issues.”