Matthew Barlocker

fix reporting of dead upstream statsd metrics

5

we have statsd servers sending stats to other statsd processes. when the upstream servers die, we don't want to persist the metrics.

Currently, the inputs on the dashboard will be remembered for each dashboard. It's not so intuitive. Perhaps have memory buttons? Or something else that is easier? not sure what to...

The legend doesn't support large numbers of lines well. Show it somewhere else. Perhaps on demand?

need a button for users to clear their local storage (the dashboard information).

Timezone is ambiguous on all pages. They are currently displayed in the user's time, but that's not apparent.

Create a dynamic alert. let it propagate. Delete the dynamic alert. The propagated alerts are never cleaned up. they must be manually deleted every time.

Right now, you subscribe to an alert. This subscribes you to the warning and error levels. Sometimes, I only want the error or the warning, not both.

Add a special tag for propagated alerts. 'propagated' will do. additionally, add another tag like 'propagated-mydynamicalert' or similar. these will both help with searching and filtering.

Create the following target in a dashboard whatever.%name%.%name|something%.whatever 2 different inputs will be created, both with name. annoying right?

provide an 'other' option in every dropdown variable box on dashboards. if selected, provide an input to allow any value.