[Labs-l] Labs reliability

[Marc-André Pelletier]

On 08/31/2014 05:57 PM, Pine W wrote:
> OK. I was hoping to see a report or graph showing server uptime
> percentages. I thought there was a way to do this.

We're currently working on a Graphite/Monitoring server to service Labs,
specifically for that reason.

That said, except for some planned major work a few times in the past
year, general uptime of Labs has been pretty much continuous.  There
have been some issues, but they tend to be localized and quickly
contained.  Beta in particular has had reliability issues, but that's a
consequence of its status as an experimental testbed sometimes running
iffy code or being used for heavy testing.

Running the typical bot is one of the things most easy to be made
reliable; Tool Labs isn't subject to the general caveat that instances
can go away at any time or that they are not supported (because they
are), and there are many mechanism in place to ensure self-contained
continuous processes (like bots) are kept running.

The biggest problem really is the poor state of enduser Documentation.
That's mostly my fault - I always end up having something else displace
it in my priorities.

-- Marc