Expressway Memory Usage

[Flo.Matalis]

Hi Experts,

Expressway C & E - Version 8.10.4

We are having issues with our expressway servers. We have an SNMP configured on the servers and we keep on getting alerts with regards to High Memory Usage. When checked in Expressway via logging in to root, we still have free memory but the SNMP server is reporting that there is a high memory usage. The workaround that we do is to restart the Expressway server and the alert will be gone.

~ # free -m
            total        used   free      shared    buff/cache available
Mem:   3951      2525    81        0          1343 1367
Swap:  10240     0         10240
~ #

Assuming that our SNMP is really interpreting the correct SNMP, does expressway server 8.10.4 consume high memory usage? And also, when checked from the ESXi, we are only seeing around 450 MB as the active guest over 4 GB as the consumed memory.

[Anurag Srivastava · Accepted solution]

Hi Florence,

# Expressway/VCS has very limited support on SNMP/MIB-2.

# The memory usage on Linux system is not something can quickly check by monitoring “single” number/parameter, but commited_as is one that shows certain level of system memory usage as “reference” number.

# We do support OID .1.3.6.1.2.1.25.2.2.0, which should return total physical memory size of the system, but no further sub-entries of this OID including committed memory size on application level.

Also Rough memory usage may calculate by “memory usage” = “hrStorageUsed” (OID 1.3.6.1.2.1.25.2.3.1.6.1) / “hrStorageSize” (OID 1.3.6.1.2.1.25.2.3.1.5.1) x 100%.
# However Expressway also utilize the swap memory (10,241MB), therefore above rough memory utilisation number may not represented the actual memory usage level either.
# Currently we don’t publish any specific number that will cover “if system utilize xx% of memory or CPU” as system may access swap memory for short time of period (i.e creating large log file) that could go higher than physical memory size in total.

# However if you need some number to monitor, Committed_AS a part of meminfo (cat /proc/meminfo | grep Committed_AS) is one we usually refer to.

# This value is very depend per customer deployment due to what service and load per sec. on system itself (enable MRA, local registration with TLS instead of TCP, refresh registration with 120 sec. instead of 1800 sec., very busy for handling call in peak business hours, etc.).

However if this value is continue to increase and go higher than physical memory size, may want to restart the system to free up the memory.
But remember, because of approximately 10GB swap memory available, and Committed_AS value is sum of total memory that all of application currently requested not “in use” value, the value may go much higher than physical memory size and system may run without any problem.

 Thanks.