Improve Error history

[IzakMarais]

I have the following alert error that keeps popping up on my dashboard

I closed it yesterday, but this morning it is back. However, there are some strange things about it:

• 1n-ago : what does this mean?
• Full History : looking at the history it seems like it has been active continually for more than a month:
  
• We did have backend instability problems a month ago, but these have been sorted out. Looking at the data for the last 24h, I see none of the usual error indications. From these dashboards one can see:
  • no HBase NSREs,
  • Number op OpenTSDB connections seems reasonable
  • Bosun's open file descriptors and go routine count look stable
• Testing the alert in the rule Editor works fine.

How can I debug this further?
We are using Bosun 0.1.

[IzakMarais]

OK, so the error is likely correct in indicating back-end instability.

All that could be improved is the history. In stead of showing a single event (and a continually active error state), it would be useful to actually see each occurence of the error in the Full History.

[captncraig]

@IzakMarais We totally reworked errors in #1301. It should be easier to debug and see individual failures. Let me know if you have continued suggestions or problems.

[IzakMarais]

Thanks, the new errors view is a big improvement.