Closed keithf4 closed 5 years ago
I believe we're going to go with tail_n_mail for real-time log monitoring for errors and pgbadger for log analysis if you want to close this issue.
Delegating this to tail_n_mail instead seems more reasonable to me than stretching the node exporter to this task. Closing this as part of decluttering old Issues we're unlikely to act on.
Idea from client to track the number of FATAL, ERROR, WARNING messages in the postgres logs and feed that to prometheus for alerting & trending.
This should be doable with a custom script to node_exporter. Will likely need some minimal info in the log_line_prefix and also an option/variable to the script to let it know what the prefix is.