Cluster health state monitoring Managed Service for YTsaurus

1. In the management console, navigate to the relevant folder.
2. In the list of services, select Managed Service for YTsaurus.
3. Click the name of your cluster and select the Monitoring tab.
4. To get started with metrics, dashboards, or alerts, select Yandex Monitoring in the list of services.

The page has two tabs, Cluster and Master, with the following charts:

• Automaton thread CPU usage: CPU utilization by the automaton thread, i.e., the master primary execution thread, as a percentage.
  • TOTAL: Aggregate CPU utilization across all master nodes.
  • ms-N: CPU utilization by the automaton thread on master node N.
• User write time Time spent on user-initiated write operations.
  • root: Time spent on root user write operations.
  • robot-odin: Time spent on robot-odin user write operations.
  • scheduler: Time spent on scheduler write operations.
  • tablet_cell_changelogger: Time spent on tablet_cell_changelogger user write operations.
• User read time Time spent on user-initiated read operations.
  • tablet_cell_changelogger: Time spent on tablet_cell_changelogger user read operations.
  • root: Time spent on root user read operations.
  • operations_client: Time spent on client user read operations.
• Master memory usage: Memory usage by the master, GB.
  • ms-N: Memory usage on master node N.
• LocalRead thread CPU usage: CPU utilization by the LocalRead thread.
  • TOTAL: Aggregate CPU utilization across all master nodes as a percentage.
  • ms-N: CPU utilization by the LocalRead thread on master node N.
• User write request rate: User-initiated write request rate, writes per second.
  • root: root user write request rate.
  • robot-odin: robot-odin user write request rate.
  • scheduler: Scheduler-initiated write request rate.
  • tablet_cell_changelogger: tablet_cell_changelogger user write request rate.
• User read request rate: User-initiated read request rate, reads per second.
  • root: root user read request rate.
  • queue_agent: queue_agent user read request rate.
  • operations_client: Client user read request rate.
• Create chunk requests: Create chunk request rate, requests per second.
  • TOTAL: Total number of create chunk requests from all users.
  • root: Number of root user create chunk requests.
  • robot-odin: Number of robot-odin user create chunk requests.
  • scheduler: Number of scheduler-initiated create chunk requests.
  • tablet_cell_snapshotter: Number of tablet_cell_snapshotter user create chunk requests.
• Under Compute resources:
  • CPU distributed vs total: Total and distributed CPU resources in the cluster.
    • distributes_resources: Amount of distributed CPU resources.
    • total_cluster_resources: Total CPU resources in the cluster.
  • GPU distributed vs total: Total and distributed GPU resources in the cluster.
    • distributes_resources: Amount of distributed GPU resources.
    • total_cluster_resources: Total GPU resources in the cluster.
  • Memory distributed vs total: Total and distributed memory resources in the cluster.
    • distributes_resources: Amount of distributed memory resources.
    • total_cluster_resources: Total memory resources in the cluster.
  • Undistributed_resources: Amount of undistributed resources.
    • memory: Amount of undistributed memory resources.
    • cpu: Amount of undistributed CPU resources.
• Under Scheduler pools:
  • CPU: usage, demand, guarantee: CPU resource usage, demand, and guarantee.
    • usage: CPU resources currently in use.
    • demand: Requested CPU resources.
    • guarantee: Guaranteed CPU resources.
  • RAM: usage, demand, guarantee: RAM resource usage, demand, and guarantee.
    • usage: RAM currently in use.
    • demand: Requested RAM.
    • guarantee: Guaranteed RAM.
  • GPU: usage, demand, guarantee: GPU resource usage, demand, and guarantee.
    • usage: GPU resources currently in use.
    • demand: Requested GPU resources.
    • guarantee: Guaranteed GPU resources.
  • Running operation count: Number of running operations.
    • lightweight usage: Number of running lightweight operations.
    • usage: Number of running operations.
  • Total operation count: Total number of operations.
    limit: Maximum allowed operations.
    usage: Number of operations in use.
• Under Problems:
  • Offline nodes: Nodes in offline status.
  • Banned nodes: Nodes in banned status.
  • Decommissioned nodes: Nodes in decommissioned status.
  • Alerts: Number of active alerts about issues on YTsaurus cluster nodes.

1. In the management console, select the folder with the cluster for which you want to set up alerts.
2. In the list of services, select Monitoring.
3. Under Service dashboards, select Managed Service for YTsaurus — Cluster Resources.
4. In the chart you need, click and select Create alert.
5. If the chart displays multiple metrics, select the data query for the relevant metric and click Continue. You can learn more about the query language in this Yandex Monitoring article.
6. Set the Alarm and Warning threshold values to trigger the alert.
7. Click Create alert.

1. Create an alert.
2. Add a status metric.
3. In the alert parameters, set the alert thresholds.