Re: [HACKERS] CLUSTER command progress monitor

On Tue, Nov 21, 2017 at 12:55 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> I agree.
>
> I have been of the opinion all along that progress monitoring needs to
> report facts, not theories. The number of tuples read thus far is a
> fact, and is fine to report for whatever value it may have to someone.

That makes a lot of sense to me. I sometimes think that we're too
hesitant to expose internal information due to concerns about it being
hard to interpret. I see wait events as bucking this trend, which I
welcome. We see similar trends in the Linux kernel, with tools like
perf and BCC/eBPF now being regularly used to debug production issues.

> The number of tuples that will be read in the future is a theory, and
> as you say, progress monitoring is most likely to be used in cases
> where theory and practice ended up being very different.

You hit the nail on the head here.

It's not that these things are not difficult to interpret - the
concern itself is justified. It just needs to be weighed against the
benefit of having some instrumentation to start with. People are much
more likely to complain about obscure debug information, which makes
them feel dumb, than they are to complain about the absence of any
instrumentation, but I still think that the latter is the bigger
problem.

Besides, you don't necessarily have to understand something to act on
it. The internals of Oracle are trade secrets, but they were the first
to have wait events, I think. At least having something that you can
Google can make all the difference.

--
Peter Geoghegan