Zabbix API maintenance creation - Host not in maintenance mode

[y4nn1ck]

Hello everyone,
I've been working on some automation on a tool i built myself and i'm running into a little issue. I use a webpage to create maintenance which in turns create the VM snapshot and the Zabbix maintenance. It all works but the host in Zabbix is not getting into maintenance mode.


My php page sends information to a python script which is making the API call. Here's that portion of script that is creating the maintenance :

m = requests.post(ZABBIX_API_URL,json={
                    "jsonrpc": "2.0",
                    "method": "maintenance.create",
                    "params": {
                        "name": maintenance_name,
                        "active_since": int(active_since),
                        "active_till": active_till,
                        "tags_evaltype": 0,
                        "hostids": [f"{hostid}"],
                        "timeperiods": [
                            {
                                "start_time": int(active_since),
                                "period": period
                            }
                        ]
                    },
                    "auth": AUTHTOKEN,
                    "id": 1
                })​

When i look into Zabbix's database, my host is not update with the maintenance id and the tag ::

MariaDB [zabbixdb]> select * from hosts WHERE hostid=11856 \G
*************************** 1. row ***************************
            hostid: 11856
      proxy_hostid: NULL
              host: CNQ-Linux2204
            status: 0
     disable_until: 0
             error:
         available: 0
       errors_from: 0
        lastaccess: 0
     ipmi_authtype: -1
    ipmi_privilege: 2
     ipmi_username:
     ipmi_password:
ipmi_disable_until: 0
    ipmi_available: 0
snmp_disable_until: 0
    snmp_available: 0
maintenanceid: NULL
maintenance_status: 0
  maintenance_type: 0
  maintenance_from: 0
​

I tried updating the host through an SQL command, which works but when the maintenance is over, the host does not get updated back to the default value.

I'm under the impression that there's some underlining process that manages the host status for maintenance.

Any ideas ?

[tim.mooney]

Always include what version of Zabbix you're using. It matters.

Here's the Perl 5 code my site uses to create a one-time maintenance via the API, for Zabbix 5.0.x:

    my $json_req = {
        'jsonrpc'   => '2.0',
        'method'    => 'maintenance.create',
        'params'    => {
            'name'              => $name,
            'active_since'      => $start_time,
            'active_till'       => $end_time,
            'maintenance_type'  => 0,
            'timeperiods'       => [
                {
                    'timeperiod_type'   => 0,
                    'start_date'        => $start_time,
                    'period'            => $end_time - $start_time
                }
            ]
        },
        'auth'      => $auth_tok,
        'id'        => $json_request_id++
    };


    #
    # either $group_ref or $host_ref (or both) should contain some
    # ids.
    #
    if (!@{$group_ref} and !@{$host_ref}) {
        carp "No hostgroups or hosts specified for maintenance period\n";
        return undef();
    }

    if (@{$group_ref} and $#$group_ref >= 0) {
        $json_req->{'params'}->{'groupids'} = $group_ref;
    }

    if (@{$host_ref} and $#$host_ref >= 0) {
        $json_req->{'params'}->{'hostids'} = $host_ref;
    }

    #
    # done building the JSON structure, now call maintenance.create
    #
    my $json_response = $rpc_client->call($url, $json_req);

[y4nn1ck]

Always include what version of Zabbix you're using. It matters.

Sorry, i am using Zabbix v5.0.3

I tried adding some of the parameter you had that i did not.. But my host is still not "under maintenance".

Here's the summary of the maintenance created by my script :



Everything looks pretty good... If i make one manually, it will look exactly the same on the frontend but the host does get updated with the maintenance information.

[tim.mooney]

As you've said, the maintenance you've created via the API looks fine.

My next suggestion would be to consider updating to the latest release in the 5.0 series, currently 5.0.30. There have been many, many fixes since 5.0.3 was released, including a few fixes involving the API. I don't see anything in the release notes that specifically mentions your issue, but it still may be worth the upgrade.

[y4nn1ck]

I'm already at that specific version..

[tim.mooney]

You said you were running version 5.0.3, above, not 5.0.30.

[y4nn1ck]

Sorry my bad

[y4nn1ck]

To anyone that is interested, i finally got back to it and found a fix. The problem was how i called the "timeperiods"... Here's what worked :

 "timeperiods": [
                            {
                                "timeperiod_type" : 0,
                                "start_time": 0,
                                "day": 1,
                                "period": period
                            }​ 

Now it created the maintenance and it is able to close it after the time period... kinda odd when you look at the documentation, it says not to use the "day" parameter. Oh well... i'm happy it works.