A single point-in-time sample of per-node execution metrics. More...
#include <node_sample.hpp>
Collaboration diagram for node_sample:
Public Attributes | |
| std::chrono::system_clock::time_point | sampled_at |
| When this sample was taken (hypertable partition key). | |
| utility::uuid::tenant_id | tenant_id = utility::uuid::tenant_id::system() |
| Tenant this node belongs to. | |
| boost::uuids::uuid | host_id |
| FK to ores_compute_hosts_tbl. | |
| int | tasks_completed {0} |
| int | tasks_failed {0} |
| int | tasks_since_last {0} |
| std::int64_t | avg_task_duration_ms {0} |
| Average task duration for tasks_since_last (milliseconds). | |
| std::int64_t | max_task_duration_ms {0} |
| Peak task duration for tasks_since_last (milliseconds). | |
| std::int64_t | input_bytes_fetched {0} |
| Total input bytes fetched for tasks_since_last. | |
| std::int64_t | output_bytes_uploaded {0} |
| Total output bytes uploaded for tasks_since_last. | |
| int | seconds_since_hb {0} |
| Seconds since the last heartbeat was received for this host. | |
Detailed Description
A single point-in-time sample of per-node execution metrics.
Published by ores.compute.wrapper nodes every ~30 seconds via NATS and persisted by ores.compute.service into ores_compute_node_samples_tbl.
