Module: check_mk
Branch: master
Commit: 37b60e0ddb4fa7549e47bc9941a3957dd1a2aa98
URL:
http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=37b60e0ddb4fa7…
Author: Mathias Kettner <mk(a)mathias-kettner.de>
Date: Fri Sep 16 09:36:28 2011 +0200
Updated bug entries
---
.bugs/375 | 29 +++++++++++++++++++++++++++++
1 files changed, 29 insertions(+), 0 deletions(-)
diff --git a/.bugs/375 b/.bugs/375
new file mode 100644
index 0000000..e689505
--- /dev/null
+++ b/.bugs/375
@@ -0,0 +1,29 @@
+Title: Clusterchecks: make host name available to check
+Component: core
+Benefit: 2
+State: open
+Cost: 2
+Date: 2011-09-16 09:31:24
+Class: feature
+
+Problem: When using cluster checks, the check itself has no
+information, on which node of the cluster a certain check item
+has been found. If we would have that information, we could
+output something like "OK - running on node zbghora10".
+
+We could extend the API for checks in a way, that the check
+can *declare* that it always wants the name of the host as
+an additional (first) column in the info table. This can
+be done with the new dictionary. Checks that are cluster
+aware (ps, df, services) can then use that information and
+display the node (maybe on a configurable basis). Maybe
+we can use None as host name in case the check is not a cluster
+check. That way we omit useless data on non-clustered checks.
+
+Check_MK need then internally to conserve that information
+in case of cluster checks. The logic for the non-cluster-checks
+must not be changed or be degraded in performance.
+
+==> That way the API for the checks keeps compatible and
+we have the "running on" information, which is also usefull
+for BI.