Last modified: 2014-08-27 22:48:06 UTC
The toolserver has the following page: http://status.toolserver.org/ This allows tools to show service status (e.g. 's7 replication broken') to their users. An alternative, simpler, implementation would be to have a single status line/site notice that can be shown by all tools, which would have the same information in plain text format.
I am very much against http://status.toolserver.org/ as a means of status information, as evidenced by the page itself: According to it for example, "server: mayapple" is "OK", while in reality it is dead :-). Status information should be determined automatically and continuously by http://icinga.wmflabs.org/ (once it is working properly). We could automatically convert the Icinga status to something that can be read by tools, but as broken services on Tools are usually fixed much more quickly than on Toolserver I wonder whether your alternative wouldn't be sufficient (a file somewhere under /data/project/.system).
Tool Labs has: * Server information and maintenance messages: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools#Instances_for_this_project * Status of services: http://icinga.wmflabs.org/cgi-bin/icinga/status.cgi?hostgroup=tools&style=detail * Status of servers: http://ganglia.wmflabs.org/latest/?c=tools * Status of the job execution grid: http://tools.wmflabs.org/?status So I'd say this bug is obsolete and a separate status keeping wouldn't be very useful. However having a single dashboard that aggregates this in a more scannable fashion would be useful (aggregated from the above sources).
(removing from "missing feature" tracking bug; this does not block any tool migration)
This will be made obsolete by labmon.
