Brocade Monitoring and Alerting Policy Suite Administrators Guide (Supporting Fabric OS v7.3.0) - Manual (Page 85)

Additional MAPS Features
● Fabric performance monitoring using MAPS.................................................................. 85
● Scalability limit monitoring...............................................................................................88
● MAPS Service Availability Module.................................................................................. 92
● Brocade 7840 FCIP monitoring using MAPS.................................................................. 94
Fabric performance monitoring using MAPS
MAPS allows you to monitor your fabrics for performance impacts, including timeouts, latency, and
throughput.
There are many distinct elements and layers in a fabric (applications, servers, switches, targets, LUNs,
and so on) and consequently multiple places that could possible be the cause of fabrics performance
impacts (bottlenecks). As each application’s behavior is unique, the impact of a bottleneck on one
individual application might be different from its impact on another application. Each MAPS event needs
to be viewed in conjunction with other server or application events to determine the actual root cause of
the problem.
MAPS timeout monitoring
MAPS monitors for timeouts on individual ports and when a timeout is seen on a port, the bottleneck
state of that port is changed to IO_FRAME_LOSS. This is then reported to the MAPS dashboard.
MAPS latency monitoring
MAPS latency detection is based on data retrieved from the port on the switch (which is just one
element in the fabric) and uses this to determine the potential impact to other flows using the fabric.
MAPS monitors the current latency on F_Ports over different time windows to determine the impact of
latency on the fabric. If it determines the latencies on these ports are severe enough to significantly
impact fabric performance, the state of that port is changed to IO_PERF_IMPACT, and the state
change is reported to the MAPS dashboard. MAPS monitors for the fabric impact state of individual
F_Ports (not on F_Port trunks) on both individual switches and Access Gateways. On an Access
Gateway set up with F_Port trunks, fabric performance monitoring is done only on the F_Ports actually
present on the Access Gateway.
MAPS and bottleneck detection
The Fabric OS bottleneck daemon supports the legacy bottleneck monitoring feature responsible for
detecting persistent bottlenecks and providing notifications. For the bottleneck daemon to produce
notifications, the sub-second bottleneck monitoring parameters must be correctly configured and the
bottlenecks must be seen persistently on the ports. If you do not use Fabric Performance Impact
monitoring in MAPS, you can continue use legacy bottleneck monitoring features in Fabric OS 7.3.0.
Refer to
Bottleneck detection with the MAPS dashboard on page 86 for additional details. All
bottleneck configurations must be made using bottleneckMon commands. Refer to the “Bottleneck
Detection” chapter in the Fabric OS Administrator's Guide for specific command details and bottleneck
monitoring parameters.
Monitoring and Alerting Policy Suite Administrator's Guide 85
53-1003147-01