Mesos Master
Mesos master exposes metrics via HTTP/Json out of the box, so all you need is just to symlink checks_available/check_mesos_master.py to checks_enabled and restart Agent.
Install
cd ${OE_AGENT_HOME}/checks_enabled
ln -s ../checks_available/check_mesos_master.py ./
Configure
At most of cases there is no need to configure Agent, but if you have non default installation of Mesos,
or if you need to monitor Mesos Master, which is not running locally for Agent node,
edit conf/bigdata.ini and make your changes at Mesos-Master section.
[Mesos-Master]
stats: http://127.0.0.1:5050/metrics/snapshot
Restart
${OE_AGENT_HOME}/oddeye.sh restart
Provides
| Name | Description | Type | Unit |
|---|---|---|---|
| mesos_allocator_mesos_allocation_run_ms | Duration of running mesos allocations | gauge | Milliseconds |
| mesos_allocator_mesos_allocation_run_ms_p99 | 99 Percentile Duration of running mesos allocations | gauge | Milliseconds |
| mesos_master_cpus_percent | Percent of in Use mesos master CPUs | gauge | Percent |
| mesos_master_cpus_revocable_percent | Mesos Master's CPUs revocable percentage | gauge | Percent |
| mesos_master_cpus_used | Amount used Mesos master CPUs | gauge | None |
| mesos_master_disk_percent | Mesos Master's dids usage in percent | gauge | Percent |
| mesos_master_disk_revocable_percent | Mesos Master disks revocable in percent | gauge | Percent |
| mesos_master_disk_used | Meso Master disks usage in bytes | gauge | Bytes |
| mesos_master_event_queue_dispatches | Master's dispatched event queue | gauge | None |
| mesos_master_event_queue_http_requests | Master's HTTP requests to event queue | gauge | None |
| mesos_master_event_queue_messages | Message in Masters even queue | gauge | None |
| mesos_master_frameworks_active | Amount of current active frameworks | gauge | None |
| mesos_master_gpus_percent | Percent of in Use mesos master GPUs | gauge | Percent |
| mesos_master_gpus_used | Amount used Mesos master GPUs | gauge | None |
| mesos_master_mem_percent | Masters memory usage in percent | gauge | Percent |
| mesos_master_mem_used | Masters memory usage in bytes | gauge | Bytes |
| mesos_master_messages_kill_task | Mater kill tasks | gauge | None |
| mesos_master_messages_reregister_framework | Master registering frameworks | gauge | None |
| mesos_master_slaves_connected | Amount of connected slaves | gauge | None |
| mesos_master_tasks_dropped | Amount of dropped tasks | counter | None |
| mesos_master_tasks_error | Amount of tasks with errors | counter | None |
| mesos_master_tasks_failed | Amount of failed tasks | counter | None |
| mesos_master_tasks_finished | Amount of finished tasks | counter | None |
| mesos_master_tasks_gone | Amount of gone tasks | gauge | None |
| mesos_master_tasks_lost | Amount of lost tasks | gauge | None |
| mesos_master_tasks_running | Amount of running tasks | gauge | None |
| mesos_master_tasks_staging | Amount of staging tasks | gauge | None |
| mesos_master_tasks_starting | Amount of starting tasks | gauge | None |
| mesos_master_tasks_unreachable | Amount of unreachable tasks | gauge | None |
| mesos_registrar_state_fetch_ms | Registrars state fetched in milliseconds | gauge | Milliseconds |
| mesos_registrar_state_store_ms_p99 | 99 percentile of registrars state fetched in milliseconds | gauge | Milliseconds |
Mesos Slave
Install
cd ${OE_AGENT_HOME}/checks_enabled
ln -s ../checks_available/check_mesos_slave.py ./
Configure
At most of cases there is no need to configure Agent, but if you have non default installation of Mesos,
or if you need to monitor Mesos Master, which is not running locally for Agent node,
edit conf/bigdata.ini and make your changes at Mesos-Master section.
[Mesos-Slave]
stats: http://127.0.0.1:5051/metrics/snapshot
Restart
${OE_AGENT_HOME}/oddeye.sh restart
Provides
| Name | Description | Type | Unit |
|---|---|---|---|
| mesos_slave_cpus_percent | Slaves CPU usage in percent | gauge | Percent |
| mesos_slave_cpus_revocable_used | Slaves CPU revocable usage | gauge | None |
| mesos_slave_cpus_used | Slave CPUs used | gauge | None |
| mesos_slave_disk_percent | Slave disks usage percentage | gauge | Percent |
| mesos_slave_disk_used | Slave disks usage bytes | gauge | Bytes |
| mesos_slave_executors_running | Amount of running executors | gauge | None |
| mesos_slave_executors_terminated | Amount of terminated executors | counter | None |
| mesos_slave_executors_terminating | Amount of terminating executors | gauge | None |
| mesos_slave_frameworks_active | Active frameworks on current node | gauge | None |
| mesos_slave_gpus_revocable_used | Slaves GPU revocable usage | gauge | None |
| mesos_slave_gpus_used | Slaves GPU usage in percent | gauge | Percent |
| mesos_slave_invalid_status_updates | Invalid status update | gauge | None |
| mesos_slave_mem_percent | Slave used memory percentage | gauge | Percent |
| mesos_slave_mem_revocable_percent | Slave memory revocable percentage | gauge | Percent |
| mesos_slave_mem_used | Slave used memory in bytes | gauge | Bytes |
| mesos_slave_recovery_errors | Recover errors | gauge | None |
| mesos_slave_tasks_failed | Amount of failed tasks on current node | gauge | None |
| mesos_slave_tasks_killing | Amount of failed killing on current node | gauge | None |
| mesos_slave_tasks_lost | Amount of lost killed on current node | gauge | None |
| mesos_slave_tasks_running | Amount of running tasks on current node | counter | None |