Monitoring Vector

This document will cover monitoring Vector.


Vector writes all output to STDOUT, therefore, you have complete control of the output destination. Accessing the logs depends on your service manager:


If you are not using a service manager, and you're redirecting Vector's output to a file then you can use a utility like tail to access your logs:

tail /var/log/vector.log

Tail logs:

sudo journalctl -fu vector

Tail logs:

tail -f /var/log/vector.log

Tail logs:

tail -f /usr/local/var/log/vector.log


By default, Vector logs on the info level, you can change the level through a variety of methods:



-v flag

Drops the log level to debug.

-vv flag

Drops the log level to trace.

-q flag

Raises the log level to warn.

-qq flag

Raises the log level to error.

LOG=<level> env var

Set the log level. Must be one of trace, debug, info, warn, error.

Full Backtraces

You can enable full error backtraces by setting the RUST_BACKTRACE=full env var. More on this in the Troubleshooting guide.

Rate Limiting

Vector rate limits log events in the hot path. This is to your benefit as it allows you to get granular insight without the risk of saturating IO and disrupting service. The tradeoff is that repetitive logs will not be logged.


Currently, Vector does not expose Metrics. Issue #230 represents work to run internal Vector metrics through Vector's pipeline. Allowing you to define internal metrocs as a source and then define one of many metrics sinks to collect those metrics, just as you would metrics from any other source.


Please refer to our troubleshooting guide: