Lewis John McGibbney created NUTCH-3239:
-------------------------------------------
Summary: Add a counter-name migration guide for the NUTCH-3132
rename
Key: NUTCH-3239
URL: https://issues.apache.org/jira/browse/NUTCH-3239
Project: Nutch
Issue Type: Sub-task
Reporter: Lewis John McGibbney
[NUTCH-3132|https://issues.apache.org/jira/browse/NUTCH-3132] (commit
f71bab402) replaced the previous counter group and counter names with the
Prometheus-style names in \{{org.apache.nutch.metrics.NutchMetrics}}. Anyone
with a dashboard, alert, or script built on the old names has no mapping from
the old name to the new one. The [metrics
table|https://cwiki.apache.org/confluence/display/NUTCH/Metrics#Metrics-MetricsTable]
on the Nutch wiki still documents the old names.
h2. What changed
The break is both the group and the counter name. These rows were checked
against the source before that commit and against the current constants:
||Old group||New group||
|FetcherStatus|nutch_fetcher|
|Generator|nutch_generator|
|IndexerStatus|nutch_indexer|
|CrawlDB status|nutch_crawldb|
|CrawlDB filter|nutch_crawldb_filter|
|injector|nutch_injector|
|UpdateHostDb|nutch_hostdb|
|ParserStatus|nutch_parser|
|DeduplicationJobStatus|nutch_dedup|
|CleaningJobStatus|nutch_cleaning|
|WebGraph.outlinks|nutch_webgraph|
||Old counter||New counter||
|bytes_downloaded|bytes_downloaded_total|
|robots_denied|robots_denied_total|
|URL_FILTERS_REJECTED|url_filters_rejected_total|
|urls_injected|urls_injected_total|
|indexed (add/update)|indexed_total|
Some counters were removed rather than renamed. Examples from the same commit:
\{{Generator}} / \{{URL_FILTER_EXCEPTION}}, \{{Generator}} /
\{{MALFORMED_URL}}, and \{{IndexerStatus}} / \{{errors (ScoringFilter)}}. Those
failures are now \{{ErrorTracker}} category counters.
h2. Proposal
Publish one old-to-new table covering every counter group and counter name that
NUTCH-3132 changed. Take the old names from the parent of commit f71bab402, and
check each new name against the current constants in \{{NutchMetrics}}.
Publish it in both places an operator looks:
* the Nutch wiki Metrics page, above the existing metrics table, with a note
that the new names are in use from 1.22 (\{{NutchMetrics}} is annotated
\{{@since 1.22}})
* the release notes for the release that contains NUTCH-3132
h2. Acceptance
* Every renamed group and counter appears once, with its old name and its
current \{{NutchMetrics}} value.
* Counters that were removed rather than renamed are listed separately,
including the three examples above, so a missing series can be told apart from
a renamed one.
* The wiki page states which Nutch version introduced the new names.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
