Good Afternoon
We are setting up a large deployment, looking to go to 40 to 50 proxies. We are now in a pilot testing mode and have setup a 4.4.1 server with three 4.4.1 proxies.
We have a two web servers 4 cpu/8GB ram, two zabbix servers 4 cpu/8gb ram ( these are in a HA pacemaker so we only use one at a time) and a MySQL 8.0.18 cluster database 8cpu/32gb ram each, (I know the cluster is not what zabbix likes but it was a requirement from my management)
The proxies are 4cpu/8gb MySQL 5.7.x database.
Looking at the process graphs I do not see anything that stands out like 100% processes or high cache usage.
I do see ~2% WaitIO on the database servers that seems a bit high.
The matrix that I am concerned about is the backlog on the proxy 3
proxies
proxy host count item count vps
proxy 1 9 944 1.85
proxy 2 23 23 6.95
proxy 3 512 52935 96.9
queue 5s 10s 30s 1m 5m 10m
proxy 1 0 3 5 11
proxy 2 3
proxy 3 454 1761 1 136 10 722
server 2
System Information
host 719 557/15/147
items 60165 54927/1870/3368
triggers 18766 7219/11547/21/7198
users on line 2 <--- this will go way up when we go live I would think
vps 109.67
From Server
############ ADVANCED PARAMETERS ################
StartPollers=30
# StartIPMIPollers=0
StartPreprocessors=8
StartPollersUnreachable=2
StartTrappers=80
StartPingers=8
StartDiscoverers=20
StartHTTPPollers=8
# StartTimers=1
# StartEscalators=1
StartAlerters=12
# JavaGateway=
# JavaGatewayPort=10052
# StartJavaPollers=0
# StartVMwareCollectors=0
# VMwareFrequency=60
# VMwarePerfFrequency=60
# VMwareCacheSize=8M
# VMwareTimeout=10
# SNMPTrapperFile=/tmp/zabbix_traps.tmp
SNMPTrapperFile=/var/log/snmptrap/snmptrap.log
# StartSNMPTrapper=0
# ListenIP=0.0.0.0
# ListenIP=127.0.0.1
# HousekeepingFrequency=1
# MaxHousekeeperDelete=5000
CacheSize=512M
CacheUpdateFrequency=120
StartDBSyncers=6
HistoryCacheSize=1G
HistoryIndexCacheSize=256M
TrendCacheSize=512M
ValueCacheSize=512M
Timeout=30
# TrapperTimeout=300
# UnreachablePeriod=45
# UnavailableDelay=60
# UnreachableDelay=15
AlertScriptsPath=/usr/lib/zabbix/alertscripts
ExternalScripts=/usr/lib/zabbix/externalscripts
######### PROXY SPECIFIC PARAMETERS #############
# ProxyLocalBuffer=0
ProxyOfflineBuffer=4
HeartbeatFrequency=180
# ConfigFrequency=3600
DataSenderFrequency=30
############ ADVANCED PARAMETERS ################
StartPollers=100
# StartIPMIPollers=0
StartPreprocessors=8
StartPollersUnreachable=20
StartTrappers=24
StartPingers=10
StartDiscoverers=8
StartHTTPPollers=4
# JavaGateway=
# JavaGatewayPort=10052
# StartJavaPollers=0
# StartVMwareCollectors=0
# VMwareFrequency=60
# VMwarePerfFrequency=60
# VMwareCacheSize=8M
# VMwareTimeout=10
SNMPTrapperFile=/var/log/snmptrap/snmptrap.log
# StartSNMPTrapper=0
# ListenIP=0.0.0.0
# HousekeepingFrequency=1
CacheSize=256M
# StartDBSyncers=4
HistoryCacheSize=1024M
HistoryIndexCacheSize=64M
Timeout=30
# TrapperTimeout=300
UnreachablePeriod=360
UnavailableDelay=120
UnreachableDelay=30
ExternalScripts=/usr/lib/zabbix/externalscripts
We are setting up a large deployment, looking to go to 40 to 50 proxies. We are now in a pilot testing mode and have setup a 4.4.1 server with three 4.4.1 proxies.
We have a two web servers 4 cpu/8GB ram, two zabbix servers 4 cpu/8gb ram ( these are in a HA pacemaker so we only use one at a time) and a MySQL 8.0.18 cluster database 8cpu/32gb ram each, (I know the cluster is not what zabbix likes but it was a requirement from my management)
The proxies are 4cpu/8gb MySQL 5.7.x database.
Looking at the process graphs I do not see anything that stands out like 100% processes or high cache usage.
I do see ~2% WaitIO on the database servers that seems a bit high.
The matrix that I am concerned about is the backlog on the proxy 3
proxies
proxy host count item count vps
proxy 1 9 944 1.85
proxy 2 23 23 6.95
proxy 3 512 52935 96.9
queue 5s 10s 30s 1m 5m 10m
proxy 1 0 3 5 11
proxy 2 3
proxy 3 454 1761 1 136 10 722
server 2
System Information
host 719 557/15/147
items 60165 54927/1870/3368
triggers 18766 7219/11547/21/7198
users on line 2 <--- this will go way up when we go live I would think
vps 109.67
From Server
############ ADVANCED PARAMETERS ################
StartPollers=30
# StartIPMIPollers=0
StartPreprocessors=8
StartPollersUnreachable=2
StartTrappers=80
StartPingers=8
StartDiscoverers=20
StartHTTPPollers=8
# StartTimers=1
# StartEscalators=1
StartAlerters=12
# JavaGateway=
# JavaGatewayPort=10052
# StartJavaPollers=0
# StartVMwareCollectors=0
# VMwareFrequency=60
# VMwarePerfFrequency=60
# VMwareCacheSize=8M
# VMwareTimeout=10
# SNMPTrapperFile=/tmp/zabbix_traps.tmp
SNMPTrapperFile=/var/log/snmptrap/snmptrap.log
# StartSNMPTrapper=0
# ListenIP=0.0.0.0
# ListenIP=127.0.0.1
# HousekeepingFrequency=1
# MaxHousekeeperDelete=5000
CacheSize=512M
CacheUpdateFrequency=120
StartDBSyncers=6
HistoryCacheSize=1G
HistoryIndexCacheSize=256M
TrendCacheSize=512M
ValueCacheSize=512M
Timeout=30
# TrapperTimeout=300
# UnreachablePeriod=45
# UnavailableDelay=60
# UnreachableDelay=15
AlertScriptsPath=/usr/lib/zabbix/alertscripts
ExternalScripts=/usr/lib/zabbix/externalscripts
######### PROXY SPECIFIC PARAMETERS #############
# ProxyLocalBuffer=0
ProxyOfflineBuffer=4
HeartbeatFrequency=180
# ConfigFrequency=3600
DataSenderFrequency=30
############ ADVANCED PARAMETERS ################
StartPollers=100
# StartIPMIPollers=0
StartPreprocessors=8
StartPollersUnreachable=20
StartTrappers=24
StartPingers=10
StartDiscoverers=8
StartHTTPPollers=4
# JavaGateway=
# JavaGatewayPort=10052
# StartJavaPollers=0
# StartVMwareCollectors=0
# VMwareFrequency=60
# VMwarePerfFrequency=60
# VMwareCacheSize=8M
# VMwareTimeout=10
SNMPTrapperFile=/var/log/snmptrap/snmptrap.log
# StartSNMPTrapper=0
# ListenIP=0.0.0.0
# HousekeepingFrequency=1
CacheSize=256M
# StartDBSyncers=4
HistoryCacheSize=1024M
HistoryIndexCacheSize=64M
Timeout=30
# TrapperTimeout=300
UnreachablePeriod=360
UnavailableDelay=120
UnreachableDelay=30
ExternalScripts=/usr/lib/zabbix/externalscripts