When weighing up the positives and the negatives, it seems that Logstash is a better alternative to Graylog and as such, will be used instead. We need to deploy a logstash instance, and ensure that we're collecting the data from the other servers using filebeat. This ticket supersedes T29.
Description
Description
Related Objects
Related Objects
- Mentioned In
- T29: Set up and configure Graylog
- Mentioned Here
- T29: Set up and configure Graylog
Event Timeline
Comment Actions
Logstash has been deployed to logstash.telepedia.net; it is collecting logs from nginx, and system at the minute; and is collecting system metrics for MW* and HAProxy via metricbeat; it still needs to be configured for the following:
- HAProxy module for connection info;
- System for Redis, Memcached, Phabricator
- Redis module
- Memcached module.
Pretty neat information thus far and looks like resources are more than ample enough. Also needs to be set up to log to different indexes', ie, Nginx should log to nginx-*, and also will need to investigate MediaWiki monolog channels for separating MediaWiki error logs out; but that's for another day.