web: monitoring for *.golang.org properties

[andybons]

talks.golang.org was down for several hours, but there were no alerts.

We should do an inventory of our web properties and set up (at least) basic reachability checks, along with a plan for how we’d like to handle the alerts.

related issues:
#21315
#15760

/cc @bradfitz @adams-sarah

[s-mang]

cc @broady who has done the alerts for playground (and other sites?)

[andybons]

Alerts are now up for golang.org, play, and tour. Since we still run some services on the legacy <version name>.golang.org instead of putting them in their own modules, we can’t set up monitoring for those services.

We should move everything to services instead of versions of the godoc app.

EDIT: we should still move everything to services but we can still set up monitoring on those properties using the legacy GAE <version name>.golang.org.

[gopherbot]

Change https://golang.org/cl/96416 mentions this issue: devapp: start of status handler for monitoring

[gopherbot]

Change https://golang.org/cl/97516 mentions this issue: devapp: revert status changes

[andybons]

Uptime monitoring has been added for

• godoc.org
• golang.org
• blog.golang.org
• build.golang.org
• google.golang.org
• perf.golang.org
• perfdata.golang.org
• play.golang.org
• talks.golang.org
• tour.golang.org
• beta.golang.org
• maintner.golang.org
• tip.golang.org
• dev.golang.org
• farmer.golang.org
• gerritbot.golang.org
• http2.golang.org
• pubsubhelper.golang.org
• vcs-test.golang.org
• sum.golang.org
• proxy.golang.org
• index.golang.org

[gopherbot]

Change https://golang.org/cl/210740 mentions this issue: status: delete unused, incomplete status client & server