Changes

Jump to: navigation, search

Balrog

34 bytes added, 18:06, 9 April 2018
ELB Logs
== ELB Logs ==
The production instance of Balrog publishes logs to two different S3 bucketswhich are [https://sql.telemetry.mozilla.org available for querying in Redash]. The relevant tables are:* The nginx access balrog_elb_logs_aus{3,4,5} - These tables contain update request records sourced from the ELB logs (that contain all of the update requests we receivenamed domain (eg: aus5) are published . If you're looking to '''balrogdo ad-us-west-2-elb-logs'''. These logs hoc queries of update request (eg: estimate how many users are very largeon a particular version or channel), and the balrog_elb_logs_aus5 is probably what you're unlikely want to be able to download them for local querying. The best way to work with them is through Athenaquery.* The rest of balrog_elb_logs_aus_api - This table contains request logs for the logs are published to '''netaus-mozaws-prod-us-west-2-logging-balrog''', in the "firehose/s3" directoryapi.mozilla. Within that there are subdirectories for different parts of Balrog:org domain** balrog.admin.syslog.admin log_balrog_admin_nginx_access - This table contains access logs for the admin wsgi app outputsourced from nginx access logs.** balrog.admin.nginx.{access,log_balrog_admin_nginx_error - This table contains error} contain logs for the admin access & error logs app sourced from nginx. The access error logs are generally a subset of the wsgi app output (which logs requests with a bit of extra detail).** balrog.admin.log_balrog_admin_syslog_admin_fixed - This table contains syslog.agent contains output from the agent admin app output's Docker container.** balrog.admin.log_balrog_admin_syslog_agent - This table contains syslog.cron contains cronjob output (eg: from the history cleanup and production database dump)agent's Docker container.** balrog.web.log_balrog_web_syslog_web_fixed - This table contains syslog.web contains output from the public wsgi app output's Docker containers. Note that this app does _not_ log requests, so this is largely warning/exception output Redash should show you the table schemas in the pane on the left. If not, you care about requests to the public app, use the nginx access logs (see above)can inspect with them with "describe $table".
== Backups ==
Canmove, confirm
6,438
edits

Navigation menu