SNMP Traps Collector

Description

• Waits on the /opt/opsview/snmptraps/var/run/snmp fifo and once a trap is received it reads it and processes it via the traps processing lib available with the opsview-snmptraps-base.
• Once a trap is processed, it is either a result which is sent to the results forwarder fifo, or an exception which is sent to the snmp-traps-exception-queue
• This also manages traces, start and stopping of traces set on a host.
• If a trace is enabled on a host, it runs that trace for the given duration in minutes or hours, max is 48hrs.
• The collector manages traces, starting, stopping, removal and update.
• Collector reads messages from the snmptrace-request-<machine-ref> queue and sends messages to the orchestrator-queue and snmp-traps-exception-queue

Dependencies

• The package requires opsview-datastore to be installed on the collector so that it can create trace documents.
• The package requires access to the opsview-messagequeue to send traces and exceptions to the result-snmp.
• The package requires access to the opsview-messagequeue to receive trace request message from the orchestrator

Installation

This is deployed when installating Opsview Monitor.

See MIBs for SNMP Traps and Gets to add your MIBs that are required for translating SNMP Traps.

Configuration

Configuration options available for the component:

• The default configuration file is available in /opt/opsview/snmptrapscollector/etc/snmptrapscollector.defaults.yaml.

• The user can provide their configs in the same locations as above in the file named snmptrapscollector.yaml

• The defaults config file:

---
# This is the default configuration file.
# Please only make configuration changes in "snmptrapscollector.yaml".
#
snmptrapscollector:
    queue_max_size: 0
    worker_timeout_secs: 30
    snmptrapstrace_channel_timeout: 10
    orchestrator_reply_delay_secs: 10
    traps_fifo: /opt/opsview/snmptraps/var/run/snmp
    results_fifo: /opt/opsview/var/results.sock
    snmp_config_file: snmptraps.cfg
    script: /opt/opsview/snmptrapsbase/bin/snmptrap2opsview
    local_message_queue: &local_message_queue
        provider: rabbitmq
        encoder: aes
        encoder_key: ZHVtbXllbmNvZGVya2V5MTIzNA==
        host: 127.0.0.1
        port: 35672
        management_port: 45672
        user: opsview
        password: opsview
    snmptrapsexception:
        queue: snmp-traps-exception-queue
        messagequeue: *local_message_queue
        durable: true
    snmptrapstrace:
        queue: snmptraptrace-exchange:direct.snmptrace-request-@:?
        messagequeue: *local_message_queue
        durable: true
    orchestrator:
        queue: orchestrator-queue
        messagequeue: *local_message_queue
        durable: true
    registry:
        provider: etcd
        host: 127.0.0.1
        port: 12379
    datastore:
        name: opsview-collector
        cache: opsviewcache,max_age=5;max_size=10
        connection:
            provider: couchhttp
            host: 127.0.0.1
            port: 15984
            user: opsview
            password: opsview

Management

Managing the component

The command to start, stop and restart the component is:

$ sudo /opt/opsview/watchdog/bin/opsview-monit <start|stop|restart> opsview-snmptrapscollector

The component logs in the syslog like all other components. Note: You may see errors in syslog regarding No such file or directory: "/opt/opsview/snmptrapscollector/var/conf/snmptraps.cfg" until the first SNMP Trap rule is associated with a host and a Reload is performed.