This option may be used to specify a string that is expected as the query
return value. The string is compared to the query result using strcmp().
The option may not be mixed with -w/-c.
This metric checks the execution time of the specified query. In case the
query does not return any (parsable) data, this is not treated as an error
when using this metric.
The -m option may be used to specify a metric to check the thresholds against.
This is more flexible than --conntime-warning/--conntime-critical (-W/-C) as
it may be extended to support further metrics without introducing tons of
arguments. Also, it does not make much sense to check the conntime and query
result at the same time.
Currently, the metrics CONN_TIME and QUERY_RESULT are available.
A query is no longer required unless QUERY_RESULT (the default) is used.
This plugin connects to an SQL database using libdbi, thus supporting all
database backends supported by libdbi. It will then issue the specified SQL
query and check the result (the numeric value of the first column of the first
row to be precise) against the specified warning/critical ranges.
The performance data includes the connection time (µs-resolution as provided
by gettimeofday()) and the query result.
Some versions of OpenSSL fail to negotiate the SSL connection with at
least some versions of Tomcat if stateless SSL session resumption
support (see RFC4507) is enabled:
| CRITICAL - Cannot make SSL connection
| 140099330348712:error:140943F2:SSL routines:SSL3_READ_BYTES:sslv3 alert unexpected message:s3_pkt.c:1195:SSL alert number 10
The problem is reproducible with OpenSSL 1.0.0h, but not with OpenSSL
0.9.8o-4squeeze12 (as shipped with Debian 6.0.4). We work around it by
disabling the RFC4507 functionality when using OpenSSL versions which
support it.
Thanks to Dag Bakke for reporting the issue and for giving me access to
a server I could use to reproduce the problem.
Add a note to the --help output which clarifies that check_http doesn't
perform certificate verification (beyond what the "-C" option does).
(Suggested by Michael Renner in Debian bug report #644627, forwarded by
Jan Wagner.)
Fix the code which accepts a comma-separated list of labels specified
via the "-l" option.
(Spotted by Oskar Liljeblad in Debian bug report #647020, forwarded by
Jan Wagner.)
Abort immediately if we don't receive a server greeting or if the
greeting doesn't contain the "--expect"ed string (by default: "220")
instead of blindly sending the EHLO/HELO line.
Spotted by Daniel Piddock, see Debian bug report #611914.
This patch adds a check for the certificate cn (hostname) to normal
certificate checks. It returns CRITICAL if th cn is missing, otherwise it
prints it in the normal output.
Patch by Stéphane Urbanovski
1. Timetick test could fail with uptime > 115 days. Thresholds are
double type, so it's safe to put a large number even for 32bit systems.
2. Add a test based on an invalid bug report, worthy anyway.
My Test::More wouldn't print the total number of tests anymore, moving
the plan at the top appears to fix it.
At the same time I made check-http.t eval the special modules so it can
skip the tests instead of failing.
This reverts commit 896962a1ad.
Conflicts:
NEWS
plugins/tests/check_snmp.t
Notes:
Reverting because I rebased a patch that was doing the same thing, plus
fixing more related regressions, and both didn't work together.
I kept the tests intact except for one that wouldn't pass on 1.4.14
either
- Use /var/tmp for state if no state dir environment variable is set,
this avoid the need for a writable localstatedir during tests.
- Use "rm -f", mostly to avoid printing out garbage of the directory
doesn't exists
Original patch to make Timeticks works as in check_snmp v1.4.14, it turns
out is_numeric isn't so useful and treating all types as numeric works
best for backwards-compatibility. This is how it used to work in 1.4.14.
As a special case, I also make calculate_rate look up for numeric values
as it would otherwise return the last value instead.
Various values (dused_pct, dfree_pct, inodes_free, inodes_total, ...) are now carried in
the parameter_list structure. Assignments have been moved to a subroutine preventing
code redundancies.
Group metrics are now calculated based on units rather than blocks. This fixes freespace calculation
when blocksizes differ within a group.
