On Oct 11, 2007, at 14:16 , Randy Wigginton wrote:
I have modified memcached as well as the java client library; with
these modifications, and very few lines of code in the application,
I can tell precisely how many URLs and SQLs are executing on a
particular machine in any given minute or in any given hour. I can
tell you the average execution time, the maximum execution time, as
well as the number of failures. I can tell you which URLs were
expensive, which URLs invoked SQL statements, which urls failed
most often. Coupled with a small mysql database, I can give you
more operational statistics on our site than many larger sites have
available.
I do a similar thing in application with a small amount of java
code. If I had to guess, it sounds like you want to do the same kind
of thing I'm doing, but in memcached instead of in a server instance.
Mine has a couple of forms:
@TimedMethod("blah.doSomething")
public Something doSomething(...) {
// expensive stuff
}
With the above, I automatically get stats that look like this (time
units in milliseconds):
blah.doSomething = compstat: count=26345 sum=461399.000000
min=7.000000 avg=17.513722 davg=17.517460 max=1731.000000
stddev=66.180928
You can also directly use the stats and do your own timings within
code (or use a simple counter variant).
The counter variant is analogous to incr. Adding a similar data
structure to keep running timing stats would be pretty cheap and
easy, but I don't know how generally applicable it'd be. You'd lose
the general discoverability I rather like, and you'd have to modify
your key space to include the origin server somehow.
How does yours work?
--
Dustin Sallings
