emmanuxhi there
emmanuxsomeone know when 3.4 will be released?
emmanuxI mean, an approximate date... we all know that it depends on a lot of factors... but... approximately?
emmanuxToo risky installing it as it is today?
emmanuxin a production environment.
emmanux3.4 has too killer features, that I can't afford to miss in my new projects.
emmanuxConfigurable JMX endpoints
emmanuxand
emmanuxParallel processing of alerts
emmanuxToo good to not have!
emmanuxLast time that I had to monitor a JBoss instalation it was a pain, I had to make an horrible hack using the DNS field of the host to hint the zabbix-java-gateway what to connect to exactly.
emmanuxThis new version has it already solved!
emmanuxPlans to release it anytime soon?
emmanuxhttp://www.zabbix.com/life_cycle_and_release_policy says "3rd quarter, 2017"
emmanuxScrew it, I'm installing it.
emmanux4.0 September 2017? That's around the corner.
emmanuxSomething tells me that it will be delayed. I hope to be wrong.
darkl0rdHey anyone who would be able to help me with a trigger?
darkl0rdI want to manually close a trigger (disk space less than 20%) and for it to stay closed. At the moment I can close it, but it will immediately raise a new event as the situation has not been resolved.
darkl0rdSo I basically want to say: Do not raise when a) previous event was closed manually AND the the event hasn't been in a resolved state in the meantime
emmanuxdarkl0rd: good question. I don't have the answer though.
darkl0rd:)
emmanuxdarkl0rd: I don't know if zabbix differentiates between manual OK/Problem status, and normal OK/problem status. I think not.
darkl0rdThe only thing I could find, but unable to use in a trigger, is {EVENT.ACK.STATUS}
darkl0rdSo, if I could for sake of argument compare .last(#1) ...
emmanuxinteresting, you could use it, since nobody should alter the OK/Problem status without first acknowledging, as a best practice.
emmanuxMy brain is burned at this hours... and I don't have much experience with 3.2. I'm installing a 3.4 right now, and tomorrow I'm going to make some tests on it. I any idea comes to my mind, I'll let you know. Meanwhile, others here could certainly help you.
emmanuxNot at these hours, though, hehe. The russians are sleeping, I think.
emmanuxLatvia as well, hehe.
darkl0rdhaha, ok thanks though.
darkl0rdAnyone awake yet? ;-)
RedShiftsomewhat
volterHehe
zabbixbot[04ZBX-12333] 03document in known issues that "output" does not work for history.get 14created by richlv (12https://support.zabbix.com/browse/ZBX-12333)
darkl0rd:)
darkl0rdAwake enough for a question related to triggers? :P
Richlv!t ask
zabbixbotask: Do not ask for a permission to ask a question. Ask right away to save time. Be specific right away so others do not have to spend time on clarifying the details you could have provided early on.
darkl0rdI want to manually close a trigger (disk space less than 20%) and for it to stay closed. At the moment I can close it, but it will immediately raise a new event as the situation has not been resolved.
_ikke_Right, manually solving triggers was not designed for that
darkl0rdThe motivation is, this is a Windows template applied to all Windows Servers. In this specific case I'm willing ot sign off on the fact that this specific machine has less than 20% space, put in a comment in the 'close' statement - and pick it IF it would hit 15% (the next trigger).
darkl0rdIf there is another way of dealing with this, or a different process that I might be overlooking.. I'm happy to adjust my thinking here.
darkl0rdSo there is no way to check .last(#1) "TRIGGER.ACK.STATUS" or something along those lines?
_ikke_No, a trigger is not acknowledged, the event is acknowledged
_ikke_And once the event is solved, it's not longer relevant
Richlvdarkl0rd, customise the threshold with usermacrovariables instead
Richlv!t userma
zabbixbotusermacros: User macros (think of them as variables) allow to override trigger thresholds and function parameters, SNMP community and other parameters per host even if items come from a single template. See macrocontext and https://www.zabbix.com/documentation/3.2/manual/config/macros/usermacros
Richlvedit the expression to use a macro as the threshold, set the macro on the host
darkl0rdrichlv: Thanks for the suggestion
darkl0rdI do feel that a scenario in which you want to permanently surpress a "warning" type alert (acknowledge it) until a more severe trigger goes off (which cannot be manually closed) is a pretty common case?
darkl0rde.g. "20% disk free on Proxy Server", I happen to know that the chances that it will fill up further in a short period of time are minimal, therefore decide that I want to "ignore" the risk. Manually close the issue, write down my motiviation in the description. Then a day later, I was wrong and the 10% trigger goes off, this one can not be closed and requires immediate attention.
Tichowhy not just disable the useless 20% warning in the first place?
_ikke_Ticho: Then you would never get it again
TichoI mean, you're ignoring it for that particular server and for that particular filesystem, so it might as well never exist
darkl0rdTicho: on a single server - it's a template applied to like 50 others.
darkl0rdSince it's a template, I like to keep it the same. Also, it raising the alert, but me or my staff having to write down why they chose to ignore it is invaluable
darkl0rdFor instance 'Hit 20% due to an update that is downloading - will free itself up later'. Only to find a day later that it didn't - accountability.
Tichothat still sounds like an alert that doesn't need to be there, as it only wastes everyone's time
Tichoinstead, what about an alert about free space being <20% for last 24 hours, and <10% immediately?
heller__hey guys
heller__i made a discovery to search all devices wich reply to ping and have an open port 80. how can i make a agent to search the network?
_ikke_heller__: agents don't have network discovery iirc
heller__hmm oh
heller__how could i then easily setup a scenario. if i have multiple devices with port 80 open and i need to see if they are online
_ikke_I don't know how many there are, but we just would add them as hosts to zabbix
Tichoagents do not have server discovery, but servers do
Tichoerr, let me try again
Tichoagents do not have _network_ discovery, but servers do
Tichoyou can set up an ip range to scan, and various checks, such as open ports, or whether it has a running zabbix agent, ...
Tichoyou just need to have the network discovery attach a template to any such discovered host, and that template will have item and trigger prototypes to check if they're online
heller_care you explain more?
heller_if i have an agent on a new location, can i have it automatically scan the defined network and start keeping track of devices with port 80 open?
_ikke_heller_: No, the agent does not have network discovery. It's the server that does the network discovery
_ikke_heller_: You *can* however create a custom discovery script which the agent runs
_ikke_(lld discovery)
_ikke_But that would associate the checks to a single hsot
_ikke_host
darkl0rdDisconnected it seems; just read your reply.
darkl0rdActually, that's not the point - because maybe on the next incident - it is important and should be actioned immediately. The decision may have been made based on time constraints, or simply someone on the department saying that he is the cause of the issue and that it will be resolved when he's done or whatever.
darkl0rdI can think of a plethora of reasons. The main thing is, that sometimes you want an alert for a specific host to go away - which is, I figure the entire reason you guys implemented the manual close in the first place.
darkl0rdSo if one can manual close an issue, it would also be nice for that issue to not re-appear.
_ikke_darkl0rd: The only thing closing an issue do is artifically creating a solved event
_ikke_The next time the data comes in, the normal trigger logic happens
darkl0rdI understand that
darkl0rdTherefore, as per my original request.. I'm askign whether it's possible to check against the previous event whether it was manually closed.
darkl0rdThat way I could say: if trigger state equals previous trigger state and previous trigger state was manually closed: ignore
_ikke_No, you cannot do that
darkl0rdDoes it sound like a sensible enhancement request which is likely to be implemented?
_ikke_I don't know
darkl0rdI'll raise an enhancement request and see :)
Tichodarkl0rd: you can have a trigger fire or not depending on status of another trigger - perhaps that's something worth looking into here
_ikke_Ticho: There are trigger dependencies, but that's for preventing triggers to run
Tichoisn't that exactly what he wanted?
Tichofor a trigger to not run temporarily under certain conditions
zabbixbot[04ZBX-12334] 03IPMI Monitoring Returning Incorrect Result 14created by Leandro Moreira14/lmoreira (12https://support.zabbix.com/browse/ZBX-12334)
emmanuxHi guys...
emmanuxI'm trying to compile proxy and agent with the "--enable-static" option, to no avail.
emmanuxconfigure: error: Not found libxml2 library
emmanuxHowever, it compiles OK without "--enable-static".
emmanuxWHYYYY
emmanuxD-';
emmanuxshould I specify the path to something in --with-libxml2=... ?
emmanuxhttps://www.zabbix.com/forum/showthread.php?t=43566
emmanux2013...
emmanux:(
emmanuxIf I get rid of that option, the same happens with the odbc lib, and others.
emmanuxI mean, it's not a libxml2 specific problem. It has more to do with static linking, apparently.
emmanuxI'm compiling from 3.4 sources.