glebRichlv: "Status of Zabbix branches" does not feel well :(
mino__Hi everyone, when i have hosts connected via a proxy, do the icmp pings get set from the proxy or the server when i use the icmp ping template?
ayeceefrom the proxy
mino__Ok, when i want the pings to get sent from the server, do i need to create another host?
ayeceei think so
volterThat's the common solution, I guess, yes.
trikkehi trying to read a webpage with json with item but it returns http headers so json preprocessing fails anyone with a solution for this ?
q1xOskars: <- 404 :(
q1xOskars: same for other past events
Oskarstnx. Will look into it. We're still finishing up with content migration and sadly there are still some blanks in content
q1xOskars: Migrating websites is always a pain :)
Oskarsthx for understanding, Raymond. Have a fun second half of the week :)
zabbixbot[04ZBX-13452] 03Elasticsearch history storage value types do not match what is set as default from the server side 14created by Viktors Tjarve14/viktors.tjarve (12
zabbixZBX-13452 Open Updated today -- Elasticsearch history storage value types do not match what is set as default from the server side
duncan9562Morning all, does Zabbix have a way, or a view whereby you can see a list of all items that are unsupported? I understand that you can view this in the item view for a host, but as an example...I have applied a template to some 500 hosts, and a good chunk of those report unsupported can I view those in say..a list? Or is that not even possible :) - Thanks in advance.
duncan9562I looked at using the nodata trigger function, however I believe that this will not trigger when the item is not supported - correct me if I'm wrong!
_NiCCan a 3.4 agent talk to a 2.4 server or proxy at all?
volter!t compa
zabbixbotcompatibility: Zabbix server and proxies must be same major version (2.4.2 and 2.4.4 is ok), ZBXNEXT-2557 asks for proxy backward support; all older agent versions are supported down to 1.0; agents more recent than server/proxy are not supported. See
_NiCsounds like a no then
volterIt's a clear yes.
_NiC"agents more recent than server/proxy are not supported."
volterOh, right, that's what it says there.
volterIt works nevertheless.
_NiCWe will give it a try.
ExancillatusHi guys, smee again :) I have this error on a Debian 8.10 (zabbix 3.2) : Failed to fetch Unable to find expected entry 'main/binary-armhf/Packages' in Release file (Wrong sources.list entry or malformed file)
ExancillatusI'm scratching my head for over an hour now. If someone has a hint, I'll take it !
volterAre you sure there are ARM builds for 3.2?
volterBecause I doubt it.
Exancill_sorry i've been disconnected for 60s, did I miss an answer ?
_ikke_Exancill_: volter doubted there were ARM builds for 3.2
Exancill__ikke_: Thanks !
Exancill__ikke_: But then how could I have this error ? My cpu is no ARM
_ikke_Exancill_: What repo did you add?
Exancill_deb jessie main
Exancill_deb-src jessie main
Exancill__ikke_: seems legit to me
_NiClooks correct yes
_NiCUnless you need source packages, you can remove the deb-src though
Exancill__ikke_: Server is up to date apart from that. I can remove src for the test
Exancill__ikke_: Same thing
q1xExancill_: zbx 3.2 is deprecated btw
Exancill_q1x: Yes but I haven't dived into a migration process. Is it a big deal ? Do I have to go to all of my monitored machines ?
Dave_hello i need help to configurate a trigger
zabbixHello Dave_, and welcome to #zabbix. I'm your friendly house bot :) Please be patient when it takes a while for people to reply. See our channel guidelines on and have a nice stay!
q1xOskars: <- this is not correct, 3.4 is mentioned as an LTS release in that table
q1xExancill_: I'm still on 3.0 (for this exact reason :)
Exancill_q1x: Ouch it's not reassuring :/
_ikke_Exancill_: we just migrated from 3.2 to 3.4, without too much trouble
_ikke_Just the server the requires a bit more memory
_ikke_and make sure all proxies are updated as well
Exancill__ikke_: Did you have to manually update all of the agents on your monitored machines ?
_ikke_agents are forwards compattible
Exancill__ikke_: Ah ? Is it just a question of changing from 3.2 to 3.4 in the repos ?
_ikke_basically yes
_ikke_make sure you have a proper backup of the db
_ikke_in case there goes something wrong with the db migration
Exancill__ikke_: yep ! I'm going to try it then. Thanks.
Exancill_So no idea for my current itching ?
_ikke_Exancill_: We add [arch=<arch>] to the repo definiton
_ikke_deb [arch=amd64] <url>
Exancill__ikke_: Trying now
Exancill__ikke_: It works, jeeez a big thank you !
Exancill__ikke_: One more thing learned today :)
_ikke_I think we ran into it with x86
Exancill__ikke_: Good to know. An IT guy from France owes you one more :)
Dave_how can I create a trigger that warns you when went down after 15 minutes?
_ikke_.min(15m)}>1 (IIRC, 2 is down)
Dave_ok now i try it
Dave_ first I tried with net.if.out [igb3] .nodata (15m)} = 1 it worked but immediately after the status "problem" came back on "ok"
_ikke_You don't want to work with nodata()
_ikke_Do you want to check the amount of traffic, or if the interface is up/down?
Dave_i want that check if interface is down after 15 min
_ikke_Then you want a different item
Dave_the item is a network discovery interface
Dave_i just need know if it possible do it: i want receive an email when the status of that or .out go down
Dave_after a period of time / net.if.out report the amount of traffic on that interface, not whether they are up or down
Dave_hmm then i can create a trigger when there is no traffic after 15 min i receive an email
_NiCDave_, changing your requirements sounds like a bad approach
_NiCIf there's no built-in item for it, I'd look into making a userparameter with ethtool or similar
Oskarsq1x - I solved the lifecycle page misunderstanding
Oskarsnow it should be clear
tuxickAPI documentation mentions for host: "Technical name of the host."
tuxickwhich one is that?
ik_zelf is it already possible to calculate a sum of all discovered filesystems space used for a host?
tuxicki can assume it means "host name"
tuxickik_zelf: i can't imagine you can do that via discovery
tuxickbut you're free to create a calculated item of course
ik_zelftuxick yes, I could create a calculated item adding each and every fs for every host ..... It would be better it this could be generated. A calculated item that could sum values for key with wildcards in the name would do nicely .... Something like sum("vfs.fs.size[*,total]")
tuxickthat'd require some iterator, but sure, would be nice
tuxickwildcard sounds like a plan
tuxicklike that :)
_ADN_Do you know guys f there is a Zabbix shared to check an external web with though port 443 (I need to import the cert to establish the connection)
tuxickonce the cert is imported it's a normal web check
_ADN_but is there such feature on Zabbix
tuxickso what do you mean?
tuxickauto-import cert?
_ADN_Yep, I need to import a cert into zabbix for zabbix to connect to X web site (which requires a cert)
_ADN_like you mention a normal web query
_ADN_but that 443 need a previous SSL cert to be imported for the connection to be establish
tuxickyes, so what keeps you from importing the cert yourself?
_ADN_if not I will get error 400
_ADN_importing the cert is possible in Zabbix
_ADN_thats what I do not find anywhere
tuxickmaybe i'm confused, but doesn't sound like a zabbix issue?
_ADN_Zabbix is the one doing the connection
_ADN_so Zabbix need to pick up the SSL cert when doing that connection
tuxickso you stick the ca cert in your /etc/ssl, /etc/pki whereever your dist keeps them?
_ADN_ok so Zabbix can not do that
_ADN_just the normal OS wil do it
tuxicknot as far as i know anyway
tuxickbut who knows someone wrote something :)
q1xOskars: great!
_ADN_thx tuxick
_ADN_I will browse a bit longer in google
tuxickif i have to import cert for a site i'd prefer moving to another site anyway :)
tuxickbut i can imagine some LAN situation where it might be relevant
tuxickbut on lan you manage both ends
tuxickik_zelf: sec, i have an idea, checking
ik_zelftuxick :
zabbixZBXNEXT-1420 Open Updated 104 days ago -- Provide ability to create calculated item based on several discovered ones
tuxickik_zelf: is closer to the answer then
ik_zelftuxick ... not so sure. I need a per host solution. Aggregate can already sum that calculated item in a host group. That's a different step.
tuxickyes, but doing same for a host will solve it it a generic way
tuxickotoh allowing wildcard like you suggest seems nicer
tuxickalso more intuitive
ik_zelftuxick, I would go for a wildcard in the key from which the values to compute with .....
JimBurns83hi guys, does anyone know how the zabbix agent encrypts the connection with tls etc?
JimBurns83does it just send an encrypted string or is there a special handshake?
abiscuolaJimBurns83: It actually establish a complete TLS connection with the server. So there is also the whole TLS handshake phase
Dave_i want know how to set a trigger when the traffic is down send an email
JimBurns83abiscuola: where could I find this in the agent source?
abiscuolaI think in src/libs/zbxcrypto
Dave_plisss help
abiscuolaJimBurns83: You should find everything you are searching for there
JimBurns83thanks abiscuola :) I don't suppose you know where the agent interprets the user parameters? I'm interested to see if it could be modified to accept a wildcard in the key part.
abiscuolaJimBurns83: Never touched that part. Sorry
JimBurns83abiscuola no worries :) thanks again for the info!
zabbixKarma for abiscuola is now 8
arvin_Hi all i am new in irc i need to monitor the cisco nx9000 with zabbix 3.0.7 please suggest the template
tuxickarvin_: i suppose you could start with some generic snmp template?
tuxickor that
ascentHi guys, I'm trying to setup JMX monitoring with Glassfish, should I run the java gateway on the server or on the client ?
zabbixbot[04ZBXNEXT-4374] 03Clear widgets configuration on the type change 14created by Alexander Shubin14/HunterCRS (12
zabbixZBXNEXT-4374 Open Updated today -- Clear widgets configuration on the type change
gleb!t userpa
zabbixbotuserparameter: Any command, executed on the agent side that returns values for items. The most popular way to extend agent data collection. See
glebJimBurns83: Like so?
mino__Are web scenarios run from the host they are configured on?
_ikke_from the server or proxy
tuxickhmm why does a disabled host show up in items list? i was looking for "unsupported items"
_ikke_tuxick: items list?
tuxickyeah, the items list of any host, remove hostname and choose "unsupported"
tuxickto get a list of all unsupported items zabbix ran into
mino__Is there a way to configure how long the web scenario data is stored?
voltermino__: I believe that this is hardcoded for some odd reason, but you might be able to change it there.
volterA guy called Richlv wrote that ticket 8 years ago:
zabbixbotAvailable topics: actionlog, actionsimulator, activeagent, agent, agentactive, agentpassive, aggregateitems, alertscripts, API, appliance, application, ask, autoregistration, beer, bugreporting, calculateditems, compatibility, conference, configcache, demo, dependencies, discovery, escalations, externalchecks, flapping, forceitempolling, forecasting, fping, frontendnotifications, globalregexp, globalscripts, housekeeper,
zabbixbot hysteresis, icmppingitems, interface, internalmonitoring, inventory, items, java, jmx, linkindicators, LLD, logging, loglevel, logmonitoring, LTS, macrocontext, macros, maintenance, networkdiscovery, networkmaps, passiveagent, predictivefunctions, protocols, proxies, proxy, quoting, regexp, remotecommands, Riga, RMB, roadmap, screens, sender, ServerActive, simplechecks, svn, templatedscreens, templates, templating, time
zabbixbot-based, timeshift, triggerdependencies, triggerflapping, triggerfunctions, triggerhysteresis, triggers, troubleshooting, trunk, upgradenotes, usermacros, userparameter, valuecache, valuemapping, vmware, webmonitoring, webscenarios, whatsnew, xml, zabconf
zabbixbot[ZBXNEXT-340] allow to customise web scenario items (URL:
voltermino__: ^
ludHi, did someone here already managed to user the builtin zabbix Database Monitoring plugin to monitor oracle DB via odbc ?
JimBurns83gleb: no, more like UserParameter=key*[*],command $1 $2
JimBurns83gleb: no, more like UserParameter=key-*[*],command $1 $2
JimBurns83sorry, i edited the 2nd line in my head and not on the keyboard
glebJimBurns83: To me this is the same as UserParameter=key[*],command $1 $2
glebWith key being "key[one,two]"
JimBurns83i think there was an issue somewhere when i tried doing it that way about 3 years ago. i could be mistaken though
JimBurns83oh I remember the issue. it was actually timeout issues.
JimBurns83which is where i moved away from vbs and powershell and started on the c#, which resulted in me accidentally writing my own zabbix agent in c#
JimBurns83i'd share my work but it's a bit memory hungry (dot net) and it has a weird issue i've yet to resolve in the way of it causing an infinite loop under certain situations (out in the wild)
JimBurns83gleb: trying to go back to scratch and modify the agent with modules on windows atm, but that's easier said than done (for me anyway)
robbbbquick question - does/will the elasticsearch integration work from the proxy as well ?
volterI'm fairly sure the Proxy can't use ES. Why would it!
robbbbso the stats dont need to go back to the server
volterAnd who's gonna evaluate the data?
robbbbit would be ideal if the proxy did
robbbbotherwise i cant see how it can scale-out
volterHow well do you know ES?
robbbbwell enough
robbbbanyone on here developing the elasticsearch integration?
abiscuolarobbb. I'm the one that mostly did it
abiscuolaWhy do you think the proxy should store data directly in it?
robbbbjust so the server isn't a single point of contention
robbbbwe ideally want a scale-out rather than scale-up architecture
abiscuolaPlease clarify the difference between scale-out and scale-up?
abiscuolaI'm not that fond on terminologies
robbbbscale-out = more instances of something - i.e. lots of proxies to spread the load
robbbbscale-up = a bigger server
abiscuolaWell you meant vertical vs horizontal scaling then.
abiscuolaThe history syncer design work that one history syncer, store 1000 values per second max
abiscuolaAcross a max of 5 concurrent HTTP connections in parallel
abiscuolaWe also use the bulk APIs for that
abiscuolaIMHO, if you are concerned to spread the load
abiscuolajust use haproxy in front of the cluster
abiscuolaIt's really a matter of how the pipeline works
abiscuolaThe data is validated, tagged with a unique timestamp and stored in cache before being flushed
abiscuolaSo you have a proper cronology from the server
abiscuolaIt really doesn't matter for how the zabbix server is designed
abiscuolaMaybe,it's more important for the front-end
abiscuolaThe proxy also is meant to work on even embedded hardware.
abiscuolaAs a side note. You saw that the ES storage back-end is actually experimental right?
abiscuolaThe aer
abiscuolaYou can spread the load through proxies it's not the history synchronization that cause the server to suffer
abiscuolaIt's everything else :)
robbbbyeah elasticsearch is really good for scale out, i'm just concerned about everything first going through the zabbix server
robbbbso the data flow is (as i understand it) .. agent -> proxy -> server -> elasticsearch
robbbbideally it would be agent -> proxy -> elasticsearch
robbbbyes, i saw it's experimental.. i intend to start having a look though.. great you have done the work!
robbbbdoes it dual-feed at the moment? mysql and elasticsearch? or is it one or the other
abiscuolarobbb: trust me, _THAT_ part of the architecture will never be a prpblem for you.
abiscuolaYou can choose to feed Different types for both the back-ends
abiscuolaBetter, you can "direct" one type to es or mysql, but not both
robbbbyou think the zabbix server will never be a bottleneck?
abiscuolaLook. I sae depmoyments using proxies with almost a 1.000.000 metrics
abiscuolaAnd you can also have multiple servers, divided by datacenter, for example
abiscuolaOr subnet
abiscuolaAnd every servet with it's proxies
abiscuolaIf you deploy one HUUUUUUGE server, you are doing it wrong dude ;)
robbbbyeah exactly, thats why scale out is better
robbbbbut multiple unconnected servers isn't really scale out
abiscuolaWell. In this case you are monitoring a series of infrastructures
abiscuolaAnd you can "orchestrate" them using the zabbix apis
robbbbindeed, it's just not ideal