As for monitoring the journal or buffer utilisation, you can see their usage stats in: System > Nodes > Details
AFAIK, Graylog does not have anything built-in to send an alert for these issues (other than the alert given in the Web UI).
You could ingest the server.log file into graylog and then build alerts from those messages or you could setup an external system which pulls in metrics from the graylog api and have that perform your alerts (This is what I do.)
I’d recommend the latter, I’ve never tried ingesting the GL server.log into GL itself but also, in the event that ES becomes unavailable or message processing time is increased, you could be alerted of the issue far later than it occurs or not alerted at all.
I’d recommend the latter, I’ve never tried ingesting the GL server.log into GL itself but also, in the event that ES becomes unavailable or message processing time is increased, you could be alerted of the issue far later than it occurs or not alerted at all.
this can kill your Graylog at times when you really want to have it working … imagine your environment is logging shit load of messages, your Graylog having issues ingest to elasticsearch and logging this into the server.log … now that is ingested to Graylog and you start the wheel that will kill everything.
Of course, but before posting, I found in the documentation the location of many log files. But in the log files, I can not find the notification message “Utilization is too high” and “Uncommited messages deleted from journal”.
I want to know if it is possible to have the contents of System> Nodes> Details in a log file.