Stats Alarm States; Table 33 Stats Alarms - Juniper MEDIA FLOW CONTROLLER 2.0.4 - ADMINISTRATOR S GUIDE AND CLI Administrator's Manual

Media Flow Controller Administrator's Guide
ordered such that the short bucket has the smallest time duration (window) and smallest
maximum allowed count. The rate-limit applies to all three buckets simultaneously.
Separate counts of alarms are kept for error alarm events and clear alarm events. For each
alarm type, a single skip count is kept and is reported when the alarm event is later sent.
When an alarm event count is exceeded for any bucket, only the skipped count is incremented
(not any of the alarm event counts).
For example, if the short bucket allows 5 alarm events per 5 minutes, the sixth alarm and
above, that occur during the 5 minute window, only increment the skip count. When one of the
windows expires, the alarms can be sent again (assuming no other window and count limits
are exceeded) and the tracking starts again.
Table 33 Stats Alarms
Stats Alarm
(unless otherwise noted, default rising error threshold is 200000000 Bps; rising clear threshold is 100000000 Bps)
avg_cache_byte_rate *
avg_disk_byte_rate *
avg_origin_byte_rate *
cache_byte_rate *
connection_rate *
cpu_util_indiv
disk_byte_rate *
disk_io *
fs_mnt
http_transaction_rate *
intf_util *
memory_pct_used *
nkn_cpu_util_ave
origin_byte_rate *
* Disabled by default; all others are enabled by default.
Alarms
Total number of bytes served divided by system up time.
Total number of Bytes served from all disks divided by system up time.
Total number of Bytes fetched from Origin divided by system up time.
Total data bandwidth being served from RAM/Buffer cache.
Incoming connections per second, arrived by summing up all accepted connections
and dividing by system up time. Default rising error threshold is 20000 per sec;
default rising clear threshold is 10000 per sec.
Average CPU utilization. The units for the cpu_util_indiv alarm are hundredths of
a point of the one-minute load average. For example, setting it to 100 causes an
alarm if the one-minute load average is ever over 1.0 when it is sampled. Default
rising error threshold is 90%; default rising clear threshold is 70%.
Current total data bandwidth being served from Disk in the system.
Disk I/O (input/output) in kilobytes per second. Default rising error threshold is
5120 kilobytes per sec; default rising clear threshold is 4608 kilobytes per sec.
Percent free filesystem space. Default falling error threshold is 7% of disk space
free; default falling clear threshold is 10% of disk space free.
Number of HTTP transactions (GET requests) per second; calculated as number of
GET requests received so far divided by system up time. Default rising error
threshold is 40000 per sec; default rising clear threshold is 20000 per sec.
Network utilization (in B/ps). Default rising error threshold is 10485760 bytes per
sec; default rising clear threshold is 9437184 bytes per sec.
Percent of physical memory in current in use. Default rising error threshold is 90% of
physical memory used; default rising clear threshold is 87% of physical memory
used.
CPU utilization across all cores too high
Current total data bandwidth being served from Origin.
Media Flow Controller CLI Commands
Description
stats alarms 361