Closed GoogleCodeExporter closed 9 years ago
3gb of files for monitoring data seems a bit too much.
I suggest to use JavaMelody 1.37 or later, because of that line in the release
notes of 1.37:
"added: Reduce the possibility of storage overload by deleting automatically
the obsolete RRD files (which were not updated for the last 3 months), and
which were for requests which do not exist anymore (revision 2750 and r2753).
Note: the size of existing RRD files is fixed for ever and old .ser.gz files
are already automatically deleted after a year."
https://code.google.com/p/javamelody/wiki/ReleaseNotes
And I suggest to check in the statistics of requests in your reports, if you
have many http or sql requests which are the same, except some dynamic values
of parameters (click the header to sort the column by request name). If yes, it
would be better to aggregate these same http or sql requests by ignoring
dynamic values of parameters.
To do this aggregation of requests, you can use the "http-transform-pattern" or
"sql-transform-pattern" parameters. See the user guide:
https://code.google.com/p/javamelody/wiki/UserGuide#6._Optional_parameters
For example, you can use the "http-transform-pattern" parameter with a regular
expression such as "\d+" to remove numbers, such as an identifier, in the http
requests.
Or you can use the "sql-transform-pattern" parameter with a regular expression
such as "\'.+\'" to remove values between quotes in the sql requests.
Original comment by evernat@free.fr
on 6 Jul 2012 at 12:43
Thanks a lot for this thourough answer !!
I guess that solves my problem then :)
Original comment by nlesc...@gmail.com
on 9 Jul 2012 at 9:56
Hi,
I have added the sql-transform-pattern and set it to
(values \(.+\))|(VALUES \(.+\))|((\d|_)*)
I now have weird character appearing in my logs for sql requests:
$s$e$l$e$c$t$ $e$t$a$b$l$i$s$s$e$m$$.$p$k$$i$d$ $a$s$ $p$k$$,$
$e$t$a$b$l$i$s$s$e$m$$.$c$r$e$a$t$e$$d$a$t$e$ $a$s$ $c$r$e$a$t$e$$,$
$e$t$a$b$l$i$s$s$e$m$$.$c$r$e$a$t$e$d$$b$y$ $a$s$ $c$r$e$a$t$e$d$$,$
$e$t$a$b$l$i$s$s$e$m$$.$m$o$d$i$f$i$e$d$$b$y$ $a$s$ $m$o$d$i$f$i$e$d$$,$
$e$t$a$b$l$i$s$s$e$m$$.$m$o$d$i$f$i$e$d$$d$a$t$e$ $a$s$
$m$o$d$i$f$i$e$d$$,$ $e$t$a$b$l$i$s$s$e$m$$.$v$e$r$s$i$o$n$ $a$s$
$v$e$r$s$i$o$n$$,$ $e$t$a$b$l$i$s$s$e$m$$.$i$d$$l$a$g$a$p$e$s$ $a$s$
$i$d$$,$ $e$t$a$b$l$i$s$s$e$m$$.$s$y$n$c$h$r$o$n$i$s$a$t$i$o$n$$d$a$t$e$
$a$s$ $s$y$n$c$h$r$o$n$$,$ $e$t$a$b$l$i$s$s$e$m$$.$n$o$m$ $a$s$ $n$o$m$$
$f$r$o$m$ $u$t$i$l$i$s$a$t$e$u$r$ $u$t$i$l$i$s$a$t$e$u$$ $i$n$n$e$r$
$j$o$i$n$ $u$t$i$l$i$s$a$t$e$u$r$$e$t$a$b$l$i$s$s$e$m$e$n$t$
$e$t$a$b$l$i$s$s$e$m$$ $o$n$
$u$t$i$l$i$s$a$t$e$u$$.$p$k$$i$d$=$e$t$a$b$l$i$s$s$e$m$$.$f$k$$u$t$i$l$i$s$a$t$e
$u$r$
$i$n$n$e$r$ $j$o$i$n$ $e$t$a$b$l$i$s$s$e$m$e$n$t$ $e$t$a$b$l$i$s$s$e$m$$
$o$n$
$e$t$a$b$l$i$s$s$e$m$$.$f$k$$e$t$a$b$l$i$s$s$e$m$e$n$t$=$e$t$a$b$l$i$s$s$e$m$$.$
p$k$$i$d$
$w$h$e$r$e$
$u$t$i$l$i$s$a$t$e$u$$.$l$o$g$i$n$=$?$<http://192.168.0.133:8080/neo-web/monitor
ing?part=graph&graph=sql8592e2f9ef0d88e65b4c365bc6ecc8dd20ae5fa8>
I want to replace
- values (...)
- VALUES (...)
- and anything like 1_50_ because of hibernate weird aliases naming
with an empty string.
My regex is correct I think.
Is there anything I missed ?
Thanks a lot,
Nic
Original comment by nlesc...@gmail.com
on 9 Jul 2012 at 12:37
Yes, I also think that hibernate alias naming is weird. And why don't they use
a simple naming for the simple cases?
But the last part, and in particular the "_", seems to be the cause of your
problem with the regexp. Note that "$" is the replacement for the aggregated
parts like "values (...)".
If you remove "((\d|_)*)" and use "(values \(.+\))|(VALUES \(.+\))", then I
think it will be ok.
Original comment by evernat@free.fr
on 9 Jul 2012 at 2:17
Thanks !!
I managed to work it out playing with the regex :)
Original comment by nlesc...@gmail.com
on 9 Jul 2012 at 6:58
ok, bye
Original comment by evernat@free.fr
on 9 Jul 2012 at 9:23
Original comment by evernat@free.fr
on 9 Jul 2012 at 9:23
Original issue reported on code.google.com by
nlesc...@gmail.com
on 6 Jul 2012 at 10:13