by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    oh and hi sry :)
    keepeek-max
    @keepeek-max
    Because actual monitoring system is based on JMX plugin system for the moment... so if JMX is available it's easy to plug into it :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max :) can I ask which monitorig system it is ?
    keepeek-max
    @keepeek-max
    Zabbix
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max Idea :
    and forget forever about JMX :)
    keepeek-max
    @keepeek-max
    gravitee access management do not use ES :(
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max setup keycloak integratin for OpenID connect to your API , and Keycloak will do both prometheus and Elastic Beats
    keepeek-max
    @keepeek-max
    Thank a lot for your help, but I really need to know if JMX is possible ? If not, we will adapt actual monitoring system to use JSON :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max my thought is also that :
    • if you monitor with JMX, "inside" the JVM, just monitor the JVM not caring about what application is deployed inside of it,
    • and if it is abouit what happened or not, the application for sure GRavitee, will persist either in adatabase, or loggs, all events that needs to be monitored.
    • If any tracing info missing, I would try any soluition, but to use JMX : that is;, search into the RAM for the info.
    • If you have to go there, then what sounds logical to me , is to open a pull request, to change software source code, so that it adds those missing inofs either into logs or database
    keepeek-max
    @keepeek-max
    you right :) just monitor JVM JMX information (like RAM) is for moment what I need to have
    thank a lot !
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max alos Iam pushing you to prom and JSOn, because :
    • yes REST is the way to operate soft in cloud today, (instead of good old ssh and shell CLI commands)
    • Zabbix is probably of the "old monitoring model" : agents pushing infos to central server
    • while the new, and senseful, and industry adopted pattern is this new, extremely simple, and therefore just as scallable, pattern : monitor pulls whatever he wats, whern he wants it, and whe he cant', this means service is down so I throw an alert (as Tiem series)
    thas is what prometheus does
    that's why I think your best move is zabbix elastic search integration in your cse
    @keepeek-max my pleasure, I had so big change in y pojects since I could push gravitee in, I real mean to thank back :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @keepeek-max okay I searched a bit more to keep this :
    • Gravitee AM persists "everything" is mongodb, see ccc
    • So you can pull everything about Gravitee AM , from its mongodb, and load it into a separate, Monitoring dedicated Elastic Stack, using for example : https://github.com/phutchins/logstash-input-mongodb
    • and you can also pull elasticsearch data from the elasticsearch that belongs to gravitee stack, and load it into the separate Monitoring dedicated Elastic Stack,
    • Now you can aggragate all data in , that separate Monitoring dedicated Elastic Stack,
    • adding prometheus you also have time series
    • and you can have Grafana Dashboard with everything you can dream of about Gravitee
    icandir
    @icandir
    bbbb
    Brassely David
    @brasseld
    @keepeek-max @Jean-Baptiste-Lasselle Sorry for being late on this guys, but you just have to enable remote JMX connection on gravitee. What's the challenge ? How do you run it ?
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @brasseld Hi David, I slipped into @keepeek-max question, because I will soon also work on monitoring in and around gravitee :
    • @keepeek-max will add more infos, but I think he wants a nice dashboard where "he can see all what happens in Gravitee AM
    • see discussion above fro what I had as first thoughts on the subject, and I will read with high interest eerything gravitee team will write about this :)
    Brassely David
    @brasseld
    we are also providing prometheus integration if you want :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    :) did not want to push it forward, but was kind of sure you guys implemented the prometheus exporter :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    thansk ! :)
    oh execellent @keepeek-max look at what @brasseld gave us : https://docs.gravitee.io/pages/node/partial/metrics.html on the page you will see example of Show jvm_memory_used_bytes metrics
    Omar Belghiti Alaoui
    @omarbelghitialaoui_gitlab
    Hello guys, I am wondering how I can stop errors logging in gateway and management-api of am ?
    for example, if there's a problème with MongoDB connection it logs Evry 5 min I thing and the memory of the server go up. How to deactivate this ?
    Thanks
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    hi @omarbelghitialaoui_gitlab Just an immediate thogut about your problem (problème, sry :) ) :
    • If you run Gravitee AM , and there is a MongoDB Connection error : maybe I am wrong, but then I think your Gravitee AM, is completely usable (you can't do anything with it)
    • So if Gravitee AM keeps growing in resources usage, i htink :
      • if you don't use it, stop it. remove it. problem solved, 100% o fresources freed
      • if you do mean, to use it, then fix the connection error problem, and restart it : problem solved
    • interested in any further move of your issue, tell us :)
    Omar Belghiti Alaoui
    @omarbelghitialaoui_gitlab
    hi @Jean-Baptiste-Lasselle Since I will use Prometheus to follow what happens with gravitee-am I don't need it to log errors in the console of the container.
    it's not really an issue, just want to know if it's possible to do so that's it. :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @omarbelghitialaoui_gitlab Excuse-me, maybe I was not clear with what I meant :
    • say right now, today at 15:00 your Gravitee AM keeps logging connection error to MongoDB : then you know, that it is compeltely unsable, by you or any compoinent in your infra
    • so what you need to do, at 15:00 , is to shut down Gravitee AM, until you fixed the MongoDB connection issue, and then respawn
    • I even believe in my team I will forbid, to keep running Gravitee AM with those logs turned off, because it is a waste of money, and could even expose security exploits
    @omarbelghitialaoui_gitlab what I do to then have a test env, is that you can "squeeze" the RAM a docker container can use on your workstation, so you can still run it with the issue, to shoot your tests
    without RAM crashes execpt for one container, the Gravitee AM one
    # edit soruce code, git commit n push it 
    # taer down git pull source code updates, respawn it all
    docker-compose down --rmi all && git pull && docker-compose up -d
    until you fix
    Omar Belghiti Alaoui
    @omarbelghitialaoui_gitlab

    For now, we are in the test so we shut down and correct the issue if there is one but if it happens in production ( we will run it on production in one week I think) and since all the application is in the same VM if it crashes on the weekend it's not good hh :D.

    The solution will be to run Gravitee AM in his own VM but it will be nice if it doesn't crush the VM every time the Mongo cluster has issues. :)
    The scenario will not happen in production but we never know.
    Since Gravitee AM don't manage levels of logs like Infos errors ... it's not possible to do what I want to do.
    Thanks @Jean-Baptiste-Lasselle

    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @omarbelghitialaoui_gitlab My pleasure, Omar, d'aider une équipe française ^^ , et je fais cela pour remercier le travail del'équipe projet gravitee, pour son produit comment dire... énorme ?
    Omar Belghiti Alaoui
    @omarbelghitialaoui_gitlab
    Yes, it's very good and useful. :)
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    @omarbelghitialaoui_gitlab also, Omar, I believe that if your Gravitee AM grows in RAM usage, it s not becaue of the logging activity, but because "everything , every feature, is trying and failing to do something", so .. it all goes crazy inside. It's not log levels, eventually,n what you wish for, is that Gravitee has a "depleted" execution mode, like "mode dégradé", where it just stays still, idle, and just answers, "oh I'm sry, I just cant do anything".
    Omar Belghiti Alaoui
    @omarbelghitialaoui_gitlab
    yeah don't worry for me I work fine like that :), I have it in production on other project and work very fine we didn't get an issue
    Jean-Baptiste-Lasselle
    @Jean-Baptiste-Lasselle
    ok :) @omarbelghitialaoui_gitlab thank you for satisfying my curiosity :)
    Mihail Yonchev
    @insertmike
    image.png
    @brasseld I am ready to support to find out the culprit, it just happened for 2nd contextPath
    djo54150
    @djo54150
    Hello , I have upgrade my version 1.30.5 to 3.0.2 but I have a not bloking error with gateway. This error appears in gravitee.log:
    00:21:43.245 [graviteeio-node] [] ERROR i.g.p.c.i.PluginContextFactoryImpl - Unable to refresh plugin context
    org.springframework.beans.factory.BeanCreationException: Error creating bean with name 'io.gravitee.reporter.file.FileReporter': Instantiation of bean failed; nested exception is java.lang.ExceptionInInitializerError
    [...]
    Caused by: java.lang.NullPointerException: null
    00:21:43.245 [graviteeio-node] [] ERROR i.g.n.r.plugin.ReporterPluginHandler - Unexpected error while creating file
    have you a idea to solve this issue?
    Yaz
    @yazalulloa_twitter
    Is there a way to create a OAuth2 resource for all APIs?
    Brassely David
    @brasseld
    @yazalulloa_twitter not yet
    @djo54150 please provide complete stack trace