[
https://issues.apache.org/jira/browse/NUTCH-3146?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Work on NUTCH-3146 started by Lewis John McGibbney.
---------------------------------------------------
> Add Resource Utilization Metrics for Fetcher
> --------------------------------------------
>
> Key: NUTCH-3146
> URL: https://issues.apache.org/jira/browse/NUTCH-3146
> Project: Nutch
> Issue Type: Sub-task
> Components: metrics
> Reporter: Lewis John McGibbney
> Assignee: Lewis John McGibbney
> Priority: Major
> Fix For: 1.22
>
>
> This task concerns adding resource utilization metrics to the Fetcher to
> improve observability of thread utilization, queue depths, and wait times
> during crawl jobs.
> h3. Proposed Metrics
> Thread Utilization:
> * threads_active_max - Peak concurrent active threads
> * threads_spin_waiting_max - Peak threads waiting for work
> * threads_spin_waiting_sum_ms - Total spin-wait time
> Queue Depth:
> * queues_total_size_max - Peak URLs queued
> * queues_count_max - Peak number of host/domain queues
> * queues_blocked_max - Peak queues blocked due to exceptions
> * queues_in_progress_max - Peak in-flight fetches
> Wait Time:
> * queue_wait_time - Latency percentiles (p50/p95/p99) for queue acquisition
> These metrics enable Nutch admins to:
> * Identify thread contention and queue bottlenecks
> * Tune fetcher.threads.fetch and fetcher.threads.per.queue settings
> * Detect capacity issues across crawl jobs
> * Compare performance between different configurations
> Note, these metrics are designed for post-hoc analysis rather than real-time
> monitoring.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
