I would like to propose a patch to improve this situation.
Data collection should use a SiblingColumnStorageStrategy storage strategy and a PersistRegexSelectorStrategy selector strategy to filter out uninteresting rows
Here's an example of such a jnxOperatingTable for a 2-nodes SRX240H cluster:
Unfortunatelly I can't get the PersistRegexSelectorStrategy working. And, none of my currently activated datacollection make use of this selector stragy. So, I can't compare this one with existing, working, ones.
Also, PersistRegexSelectorStrategy is missing some logging. That would help if someone could add the following line in PersistRegexSelectorStrategy.visitAttribute():
@Override public void visitAttribute(CollectionAttribute attribute) { if (StringAttributeType.supportsType(attribute.getType())) context.setVariable(attribute.getName(), attribute.getStringValue()); +++ LOG.debug("visitAttribute: storing value {} for attribute named {}", attribute.getStringValue(), attribute.getName()); }
This is something done in ThresholdingVisitor.visitAttribute()
Collection of data from the jnxOperatingTable (OID's under .1.3.6.1.4.1.2636.3.1.13 - see JUNIPER-MIB at http://www.juniper.net/techpubs/en_US/junos12.1/topics/reference/mibs/mib-jnx-chassis.txt for more details -) in juniper.xml isn't made correctly.
I would like to propose a patch to improve this situation.
Data collection should use a SiblingColumnStorageStrategy storage strategy and a PersistRegexSelectorStrategy selector strategy to filter out uninteresting rows
Here's an example of such a jnxOperatingTable for a 2-nodes SRX240H cluster:
See also http://issues.opennms.org/browse/HZN-367