ManikantaHai everyone
ManikantaI have some doubts could you please clarify
ManikantaI am using latest version of zabbix (which includes the separate namespace for zabbix rather than using haproxy namespace)
ManikantaI am able to reach public_ips and internal_ips but I am not able to reach haproxy vip
ManikantaI have stucked in this stage
Manikantacould you please suggest me how to proceed with?
pidzerowho's currently the crypto expert at zabbix llc?
glebpidzero: Andris_zbx, as usual
pidzerogleb, thx! could have changed meanwhile :)
Andris_zbxpidzero: hi!
pidzeroAndris_zbx, hi! :) in case you can spare some time, i'd be very interested in your opinion about
pidzeronothing urgent though - just out of curiousity :)
pidzeromore explainations can be found in the related comment in the ticket
Andris_zbxpidzero: which ZBX does it belong to ?
pidzeroZBXNEXT-1660 - but the name and description might be misleading
SproGuys i'm sort of lost,
SproI've got items with unit kWh, which i don't want to be shown with the M/K/G/etc multiplier prefix, so i editted the and added 'kWh' to the $blacklist.
SproThis is working for just monitoring, but not i've added those items to be included in an action , so that they get emailed if a trigger fires... result: Rack 1A: 1.69 KkWh
SproAny idea where i can find the variable to fix that?
Andris_zbxpidzero: well, need to think about it.
pidzeroAndris_zbx, appreciate it!
pidzeroSpro, alerts are send by zabbix server and thus frontend modifications have no impact on alerts
SproThat sounds reasonable, i was googling already, but couldn't find an easy fix for it
glebSpro: src/libs/zbxserver/evalfunc.c: if (value_double < base || SUCCEED == str_in_list("%,ms,rpm,RPM", units, ','))
SproAh! nice, let me try and fix that anoying issue. Thanks gleb/pidzero ;)
glebSpro: And vote for ZBXNEXT-768!
zabbixbot[ZBXNEXT-768] allow to customise unit blacklist (URL:
pidzerowhat is yet another feature whose implementation might benefit from ZBXNEXT-2519
pidzerowhich possibly never ever happens :(
SproDone gleb ;)
zabbixbot[ZBXNEXT-2519] Move API to Zabbix Server (URL:
kottenCurious about zabbix, I have seen there are some neat VMware integration. How does it play with hyper-v monitoring?
hnsklooks a bit like you have to run powershell scripts to achieve that
hnsk(I'm not familiar with hyper-v monitoring but something like this seems to exist: )
kottenYeah, could be worth looking into
doricoany help to monitor ILO2 on ML350 ?
Richlvdorico, yes
Richlvor maybe
doricoi read something about a plugin to do it
doricoand something else about a teamplate
doricoZabbix 3.2
Richlvdorico, monitor what exactly ?
doricohardware informations , health, disk health, ecc
finalbetadorico, if a template exists, it will probably be on
Richlvis down ?
Richlvdorico, some of that you can monitor with ipmi
finalbetaRichlv, looks like it.
Richlvfinalbeta, or on, but i cannot check right now ;)
finalbetanor can I ;)
Richlvoh. is down, too
Richlvzalex, ping
Richlveverything is down :)
zalexRichlv: yes
zalexcore router on ISP has died
mybalzitchtold him not to leave his drink there
mybalzitchspeaking of drinks
mybalzitchtime for starbucks
Richlvzalex, a-ha. thanks for the update :)
SproAlways nice, dead routers... poor guys
zalexoptimistic downtime - a few hours, maximum - 1 day
SproNo redundant network? :(
RichlvSpro, i was holding back from comments about redundant equipment, mhm...
Richlvzalex, lattelecom or smth else ?
SproWell, since im working at a networking company, i couldn't let it go :)
zalexRichlv: I don't know. shared already all I'm aware of
RichlvSpro, yeah, non-redundant core router sounds like somebody saved some money where they should not have :)
Richlvzalex, cool, thank you :)
Richlvwhich probably explains no zabbix devs online :)
SproLovely those hosters indeed. In the netherlands we have Ziggo, which is spread all over the country. There network is "redundant" but still they manage to generate outages on a regular base ;)
Sprohosters/service providers *
nadleyhi guys
ganja\I have added the same template and belonging userparameters and script to three hosts. Only of them the gets new items via the low level discovery. I can't find any method to debug it, any suggestions?
ganja\It was nothing. It's solved. The discovery runs once per hour. I thought an hour had passed now, but it hadn't 3 minutes later the items were added.
zalexlooks like ISP's issue has been resolved
ganja\ /part
binary109does anyone know when the zabbix agent for windows will update for 3.2.4
_val_hi. Running mariadb here but I get unsupported Item key (zabbix 3.0)
_val_I mean I'm running Zabbix 3.x version
zabbixbot[04ZBXNEXT-3779] 03Add optional HTTP proxy support for local builds by docker-compose 14created by Marc14/okkuv9xh (12
Parabola_val_: those two things don't on the surface seem related. you'll need to explain your problem
pidzerowow, zbxnext-18 got the highest story points :D
_val_Parabola: ZBX_NOTSUPPORTED: Unsupported item key.
_val_from the zabbix master: zabbix_get -s -p 10050 -k
_val_Parabola: this works. E.g
_val_[0][root@test-mon-10:~ ]$ zabbix_get -s -p 10050 -k system.uname
_val_Linux andromeda.local 4.10.8-200.fc25.x86_64 #1 SMP Fri Mar 31 13:20:22 UTC 2017 x86_64
pidzeroACTION wonders whether zabbix might be affected by ...
_val_pidzero: I think I found some
_val_needed to add e.g UserParameter=mysql.version,mysql -V in the clients /etc/zabbix/zabbix_agentd.conf
zabbixbot[04ZBX-12022] 03IT services expands all trigger descriptions 14created by Alexey Pustovalov14/dotneft (12
zabbixbotERROR: Could not fetch issue description. Reason: you do not have the permission to see the specified issue.","login required.
volterAd zbxnext-18: Hooray, more actionsim work!
enderlessWith web scenarios, session/cookies are supposed to be saved across steps correct?
enderlessI have a scenario simulating a login, and on the step proceeding the login (via POST), Zabbix does not appear to use the login cookie/session from the previous step
enderlessChecked this manually with cURL and can replicate the steps
enderlessIs there anything that needs to be passed in the string post login?
enderlessi.e. steps are 1) Check login page (required string "Login User ID") -> Login (jdeowpBackButtonProtect=PROTECTED&User={User}&Password={Password}&Environment={Environment}&Role={Role}) -> Check login status (required string "Welcome <USER>")
enderlessi.e. steps are 1) Check login page (required string "Login User ID") -> 2. Login (jdeowpBackButtonProtect=PROTECTED&User={User}&Password={Password}&Environment={Environment}&Role={Role}) -> 3. Check login status (required string "Welcome <USER>")
enderlessSomething like that
enderlessStep #3 fails for the required string for a logged in user (Welcome <USER>), but succeeds if I replace it with the "Login User ID" string
enderlesswhich is telling me it's not using the saved session/cookie from step #2
hydruid_enderless: All cookies are preserved during the execution of a single scenario.
hydruid_I'm trying to find a example that matches your case
enderlesshydruid_, that's what I was thinking (and what documentation is saying).
enderlessUnfortunately, the step proceeding a login sequence queries a page (only available for a logged in user) and is failing
hydruid_enderless: are you sure that it is logging in?
hydruid_enderless: is it possible to create a page where a user can login that is automatically redirected to the page you want after logon since zabbix follow redirects?
hydruid_I know that doesn't answer your question but it's a possible alternative approach if you can confirm the user is actually logging in
enderlessThat is what I'm wondering. The login step is a POST sequence and does not fail, but it could be possible that it is failing silently for Zabbix
belicHey everyone... question.. What i'd like to accomplish is have is set a host-level macro, and then reference that variable in a discovery rule... seems like it's not possible.. thoughts?
volterThat should work just fine.
enderlessSeems my variables weren't set correctly. Manually inputting the POST variables seems to have gotten me logged in Zabbix.
enderlessNow, just need to logout. :)
enderlessIn a mess of JS and pop-up windows
belic@volter: when i try to reference the host-level macro, lets say it's called {#MACRO}, when I go to the discovery rule filter and enter {#MACRO} (note it starts with #, not $}, i get this error: "Incorrect filter condition macro for discovery rule"
belic@Volter: sorry, got those mixed up. host-level macros use $ and LLD filters use #..
volterenderless: Welcome to the world of web applications!
enderlessvolter, ha! Yea, not my first rodeo
enderlessUnfortunately, this is a closed/proprietary platform and I can't really get my hands on to dig in it
enderlessThe logout sequence is a GET request to a page that executes a JS Logout() script that launches another window
enderlessgood grief
pidzerohaven't followed the conversation but read 'get' and 'js' and 'logout()'... enderless possibly you're interested in ZBXNEXT-3515. if so, please consider to vote for it ;)
zabbixbot[ZBXNEXT-3515] Add embedded script support to Web scenario steps (URL:
enderlesspidzero, yup, that's pretty much what I'm looking for. Login via POST is fine, logout is JS BS
enderlessI voted for it :)
pidzerothx fellow! we seem quite alone with this demand though :)
enderlessI've been looking it to other options (like selenium or phantomjs) which is emulating a headless "browser".
enderlessTried anything else on your end?
pidzeronot yet. if i've understood them correctly, then these rum's don't scale well - since needing a real browser doing the actual job. however, might have changed meanwhile
pidzeroenderless, logout by a js function in almost any case still leads to a get/post request in the end. the question is rather, does this request have to include a variable by java script evalutate value?
pidzerohave you completely traced the communication when doing it manually with your browser?
enderlessYea, I could see scaling for a large infrastructure presenting a problem. Haven't really used either so I can't say
pidzero"when doing it manually with your browser" sounds somehow weird :D
enderlessUsing firebug and wireshark now to see if I can capture everything
enderlessIt's Oracle JD EnterpriseOne, and I can't grab any source to see what is being called/ran on the server side
pidzerohmm, haven't used firebug for quite a while. i had made good experience in the past with HttpFox
pidzerohere i'd argue the use wireshark is a bit overkill :)
pidzero*use of
enderlessNice. Haven't used httpfox, but I'll give it a go
enderlessWireshark is definitely overkill, but I can filter by POST/GET requests and see everything getting sent out
pidzeroenderless, when could not find out why a zabbix web scenario didn't worked, then in 99% it was worth the effort to implement the web scenario in a dumb batch script making a curl request per step
pidzeroit's fairly easy actually
pidzeroand you get the responses like you normally would only get when having access to the zabbix server/proxy who is doing the web checks.
pidzeroah! while talking about that, yet another feature request comes to my mind...
enderlessI've scripted the entire process with bash insofar, minus the logout sequence.
enderlessIf I can figure out how/what is getting called in the JS logout() sequence, the web scenario should work fine
pidzerothere it is: ZBXNEXT-2411
zabbixbot[ZBXNEXT-2411] Smart response capturing of failed web scenario step (URL:
volterbelic: So, all good now?
Psi-JackSo, question. I have a web check running every 60 seconds, and a trigger using .avg(10)>0.8, but it doesn't matter what happends, if 1 collected metric happens to be 1, it triggers it no matter what.
Psi-JackWhat I'm trying to do is only trigger it if it's not been 0 for 10 minutes consistently.
Psi-JackThe only thing I can come up with so far is .sum(10)<10
Psi-JackBut, I'd thought that avg() would calculate a float average, rather the retain an integer.
Psi-JackEspecially when I have 10 results with only 1 being 1, the remaining 9 being 0, and it still triggers.
volterPsi-Jack: What do you think avg(10) does?
volterFrom what you are writing, I think you want avg(#10).
Psi-JackWell, I was assuming avg(10) would average the past 10 minutes of metrics together.
Psi-JackOh... That's seconds...
Psi-JackNo wonder!
Psi-JackMan. I feel silly now. :)
volterHappens to all of us!
Psi-JackWell, hopefully this shuts up these stupid stupid server alerts.
Psi-JackWhile still alerting on real problems that last longer.
Psi-JackClient situation.. They wanted to monitor their health check endpoint which incorporates checks and sums them all up into one root result value which is what zabbix checks, and of course when their external SMTP server and OAUTH server stops working, it makes the alert trigger.
Psi-JackYeah... I keep telling them they need to remove those and monitor themselves, or we'll have to stop supporting it.