No keep-alive sent from client in over 180 seconds


#1

I got error No keep-alive sent from client in over 180 seconds in uchiwa UI and can’t stop this error.
is any idea?


#2

Usually happens when it's lost connection or desynchronized from the sensu server, etc.

Try a "service sensu-client restart" on the client.

···

On Tue, Oct 07, 2014 at 01:25:25AM -0700, kinpo Samohan wrote:

   I got error No keep-alive sent from client in over 180 seconds in
   uchiwa UI and can't stop this error.
   is any idea?

--
--------------------------------------------
Charlie Drage


GPG [FE8E 8D18] [charliedrage.com/public.key]
--------------------------------------------


#3

it works after sensu-server started.

Thank you Charlie.

2014年10月7日火曜日 17時43分48秒 UTC+9 Charlie Drage:

···

Usually happens when it’s lost connection or desynchronized from the sensu server, etc.

Try a “service sensu-client restart” on the client.

On Tue, Oct 07, 2014 at 01:25:25AM -0700, kinpo Samohan wrote:

I got error No keep-alive sent from client in over 180 seconds in
uchiwa UI and can’t stop this error.
is any idea?

Charlie Drage
charliedrage.com

GPG [FE8E 8D18] [charliedrage.com/public.key]


#4

I also meet this, when the time of sensu client is after sensu server 120s. So check it.

sensu check keepalive depend on timestamp, so it will create events, when the time is different.

Good Luck.

Lawrency Meng

在 2014年10月7日星期二UTC+8下午4时25分25秒,kinpo Samohan写道:

···

I got error No keep-alive sent from client in over 180 seconds in uchiwa UI and can’t stop this error.
is any idea?


#5

0down vote

I have a setup of sensu core(single server with muliple clients). I am facing the same issue - ‘No keepalive sent from client for 7807 seconds (>=180)’.

Tried the solution as suggested here: http://deploy.live/sensu/2017/04/01/sensu-no-keepalive-sent-from-client-for.html. by installing ntpserver.

[root@client-machine user]# ntpdate pool.ntp.org 31 Aug 14:07:50 ntpdate[14158]: the NTP socket is in use, exiting

[root@server-machine:/home/user]# ntpdate pool.ntp.org 31 Aug 14:08:04 ntpdate[29031]: the NTP socket is in use, exiting

At intermittently, we are finding keepalive shows GREEN status as expected and it goes to RED status often

Kindly advise to get rid of this problem!

···

On Tuesday, October 7, 2014 at 1:55:25 PM UTC+5:30, kinpo Samohan wrote:

I got error No keep-alive sent from client in over 180 seconds in uchiwa UI and can’t stop this error.
is any idea?


#6

Sensu Keepalives are sent from the client automatically every 20 seconds if the client does not send a result then that indicates a real problem. As others indicated this usually boils down to several types of issues:

  • a time sync issue (ntp/chrony can help with this)
  • an issue with the transport (rabbitmq is recommended)
  • connecting a (rogue) server that is using a different transport, for example if you accidentally connect your local sensu server in a container/vm to your existing sensu servers but do not connect to the same transport the server will correctly see that no keepalive events are coming in.
  • client is unable to connect to your transport

Unfortunately there is no silver bullet to make these go away as they are all real issues and are doing exactly what it is being told to do. The best advice I can give is work down the list of possible scenarios and determine whats happening and look for root cause.

If you are OK with ignoring keepalive events you can configure your keepalive handler on a per client basis to a handler that is a no-op (for example just runs /bin/true. Alternatively if you want keepalives but only after a longer period of time has passed you can adjust both warning and critical thresholds.