[Bug 39411] New: Provide a way to monitor VM revert times

https://bugs.winehq.org/show_bug.cgi?id=39411

Bug ID: 39411
           Summary: Provide a way to monitor VM revert times
           Product: Wine-Testbot
           Version: unspecified
          Hardware: x86
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: unknown
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

It would be nice to have a graph of the VM revert times to verify that
everything is working fine (or conversely to get a better sense that things are
wrong and action is needed).

Currently the way to do so is to ask Alexandre for the WineTestBot log, run a
perl script on it to get a CSV, and import the CSV in a spreadsheet to graph
the data. That's pretty cumbersome and obviously does not permit regular
monitoring.

Providing per-VM revert time graphs, maybe using something like RRDtool would
be really nice.

But then one could argue that revert times should just be low and stay that way
and that then monitoring them may not be needed.

[Bug 39411] Provide a way to monitor VM revert times

https://bugs.winehq.org/show_bug.cgi?id=39411

--- Comment #1 from Hans Leidekker  ---
(In reply to François Gouget from comment #0)
> But then one could argue that revert times should just be low and stay that
> way and that then monitoring them may not be needed.

I agree, and I believe reverts can be made fast. Currently reverting a VM means
going back to a really old image and then applying a snapshot including large
deltas, like service packs.

If we'd merge the snapshot into a new base image then the delta would be small,
it would essentially be the VM saved state and applying it should be fast.

[Bug 39411] Provide a way to monitor VM revert times

https://bugs.winehq.org/show_bug.cgi?id=39411

François Gouget  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #2 from François Gouget  ---
The TestBot now has a Munin plugin which makes it possible to see the long term
evolution of various parameters like jobs/hour, tasks/hour, VM host utilization
statistics, *revert times*, WineTest run times, etc.

So if a VM takes a long time to revert it will now be possible to see if it
happened gradually or on a specific date (e.g. after some change on that VM).
It will also be possible to see increases in a VM host utilization which would
show that rebalancing the workload accross VM hosts is needed or that it's time
to buy more hardware, etc.

As before the Munin site is on vm2 (so not publicly accessible :-( ).

[Bug 39411] Provide a way to monitor VM revert times

https://bugs.winehq.org/show_bug.cgi?id=39411

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #3 from Austin English  ---
Closing.