Now that you've deployed Vector you'll want to make sure it's healthy and performing. This guide will cover how to monitor and observe Vector.
Vector writes all logs to
STDOUT, accessing these logs is dependent on how
you installed Vector.
sudo journalctl -fu vector
What To Expect
How Vector performs is largely dependent on your configuration, hardware, data shape, and volume. No two deployments perform the same! With that said, Vector should perform well in most cases. The following sections provide general references points for what you can expect in production.
Our performance benchmarks are the best point of reference when understanding how Vector performs. This data should give you a general idea of Vector's performance under different scenarios.
Vector should be relatively high throughput. One a 4 core machine should produce throughput between 20mb/s to 100mb/s depending on your data and configuration.
Vector is mostly CPU bound and benefits from more cores and CPU efficient machines.
Vector should use very little memory. For average pipelines Vector uses around 200mb, but this can vary widely dependind on your configuraton.