30 likes | 128 Views
IT-PES Monitoring update. Maite Barroso. Summary of wishes. Re-use existing code ( Nagios probes) and integrate with what we’ve got (Lemon/LAS) Stream the alarms (type, severity) Make it easier to do simple correlations Move from the concept of “clusters” to “services”
E N D
IT-PES Monitoring update Maite Barroso
Summary of wishes • Re-use existing code (Nagios probes) and integrate with what we’ve got (Lemon/LAS) • Stream the alarms (type, severity) • Make it easier to do simple correlations • Move from the concept of “clusters” to “services” • Support easy verification of automated operations • Email is not a acceptable substitute for monitoring • Collaborate on a lego-set for service managers needing to do service-specific analytics • But not a monster tool • Make use of standard open-source tools
Progress • Nagios to SLS gateway, already used in production (MyProxy service) • Batch monitoring: • Main need: analytics; OK with existing probing of individual machines • Distributed database, Cassandra, to store batch monitoring data, and additional software modules for supporting the special service needs • ActiveMQ