I need to track very different log files for errors, success status, etc. And I need to capture the relevant indicators using Prometheus and show in Grafana + set some warnings on it. Prometheus + Grafana is fine. I already use them a lot with different exporters like node_exporter or mysql_exporter etc. Also a warning in the new Grafana 4.x works very well.
But I have a rather difficult task to find a suitable exporting program / program that could analyze the log files on the fly and extract metrics from them.
So far I have tried:
- mtail ( https://github.com/google/mtail ) - works, but the existing version cannot easily control more files - in general, it cannot link a specific mtail program (receive for analysis) to some specific log file + I donβt I can easily add the log file name to the tag
- grok_exporter ( https://github.com/fstab/grok_exporter ) - it works, but I can extract only limited information + one instance can control only one log file, which means that I will have to run more copies, exporting to more ports and tuning them in prometheus, making too many new points of failure
- fluentd prometheus exporter ( https://github.com/kazegusuri/fluent-plugin-prometheus ) - works, but it looks like I can extract only very simple indicators, and I can not do any advanced analysis of regular expressions of the string (s ) from the log file
Does anyone really have a running solution for monitoring advanced indicators from log files using "some exporter" + "Prometheus + Grafan"? Or instead of an exporter, some program from which I could get results using the Prometheus gateway. Thanks.
source share