fix(platform): verify InfluxDB _internal fields against live v1.13.0 - #7685
Open
jstirnaman wants to merge 14 commits into
Open
fix(platform): verify InfluxDB _internal fields against live v1.13.0#7685jstirnaman wants to merge 14 commits into
jstirnaman wants to merge 14 commits into
Conversation
Refresh the _internal measurement and field reference against a live InfluxDB Enterprise 1.x instance, restructure the page as an edition-aware wayfinder, and link it from both v1 monitoring pages so readers can actually find it.
## What changed
- Converted 6 Hugo `{{% note %}}` callouts to GitHub-style `> [!Note]` markdown in measurements-internal.md
- Converted 1 Hugo note callout in Enterprise v1 monitoring index
- Removed "Check out..." phrase linking to blog posts from OSS v1 server_monitoring.md
- Removed TODO comment about unnamed InfluxDB 3 Cloud edition
- Added InfluxDB 3 Cloud to the monitoring edition reference list
## Why
GitHub-style callouts are the documented standard per DOCS-CONTRIBUTING.md and provide consistent rendering across the site. Removed outdated "Check out..." phrasing per Google Developer Documentation Style Guide. Resolved TODO by including InfluxDB 3 Cloud in the documentation.
## Impact
Documentation consistency: all callouts in updated files now use modern GitHub syntax. Removed reference to external blog posts in favor of official documentation links.
## Verification
Hugo build: successful
No remaining Hugo shortcode callouts in affected files
All 7 GitHub-style callouts confirmed in place
## What changed - Removed 8 fields never corroborated by any source and confirmed absent on two live InfluxDB v1.12.4 instances (OSS and Enterprise) after fresh write and query activity: queryExecutor.queriesFailed, queryExecutor.queriesSlow, and tsm1_engine's six compactionPlanner* fields. - Reclassified shard.indexType: it's a tag (confirmed via SHOW STATS' tag set), not a field. Moved it into a tags note on the measurement intro instead of a fake field heading. - Removed an incorrect "(Enterprise only)" label from write.pointReqLocal; it's present on OSS too. - Added two field headings that existed in the doc body but were missing from the TOC: cluster.openConnections, httpd.fluxQueryRespBytes. - Fixed a TOC entry whose link text didn't match its own heading (fluxQueryDurationNs -> fluxQueryReqDurationNs). - Fixed two TOC anchors that never resolved: mixed-case #pointsWrittenOK/#valuesWrittenOK don't match Hugo's lowercased, deduped heading IDs (the httpd and write measurements each have a same-named field, so the real targets are #pointswrittenok-1 and #valueswrittenok-1). - Replaced the rendered "pending confirmation" note with a `<!-- VERIFY (live instance) -->` comment for the four fields and one measurement that remain unconfirmed, tracked in #7684, instead of shipping speculative status to readers. - Removed the stale .pr.md draft; this commit message and PR description supersede it. ## Why `httpd.valuesWrittenOK`, `write.pointsWrittenOK`, `write.subWriteDrop`, `write.valuesWrittenOK` are still documented by the Telegraf `influxdb` input plugin, so rather than deleting them outright they're flagged as unconfirmed pending engineering input (#7684) -- they may be limited to older 1.x releases. Everything else removed here had zero corroborating evidence anywhere in this repo or in Telegraf's docs, and is confirmed absent via two independent live instances. ## Impact No fields that are actually emitted by current InfluxDB 1.x releases were removed. Readers get a corrected field/tag list; the four still-uncertain fields no longer read as authoritative in the rendered page. ## Verification Live-tested against InfluxDB OSS v1.12.4 (port 8087) and InfluxDB Enterprise v1.12.4 (ports 8086/8091), both freshly started for this check: - `SHOW MEASUREMENTS` / `SHOW FIELD KEYS` against `_internal` on both instances, before and after writing line protocol data and running queries (including a deliberately failing one, to test queryExecutor.queriesFailed). - `SHOW STATS` (instant snapshot, unaffected by the ~10s _internal write interval) cross-checked against the persisted `_internal` database after waiting out that interval -- same result both ways. - `hh_database` couldn't be tested: reproducing it needs an active hinted handoff failure between cluster nodes, not reproducible with a single-node Enterprise setup. - `npx hugo --quiet --gc`, `.ci/vale/vale.sh`, and link-checker all pass on the changed file. Closes #7557.
Contributor
Vale Style Check Results
Errors (blocking)
Warnings (12)
❌ Check failed — fix 10 error(s) before merging. |
github-actions
Bot
added
product:v1
Contributor
🔗 Link Check Results — Link Check Bot❌ 113 broken link(s) found — fix them before merging.
Broken Links
Showing first 50 of 113 errors. See the workflow run for the full list.
|
| Source File | URL | Issue |
|---|---|---|
content/enterprise_influxdb/v1/administration/backup-and-restore/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/backup-and-restore/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/configure/config-data-nodes/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/configure/config-data-nodes/_index.md |
https://support.influxdata.com/ | Network error: SSL certificate not trusted. Use --insecure if site is trusted (e… |
content/enterprise_influxdb/v1/administration/monitor/external-monitor-setup/_index.md |
https://support.influxdata.com/ | Network error: SSL certificate not trusted. Use --insecure if site is trusted (e… |
content/enterprise_influxdb/v1/administration/monitor/external-monitor-setup/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/internal-vs-external/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/internal-vs-external/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/kapacitor-measurements/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/kapacitor-measurements/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/measurements-internal/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/measurements-internal/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/monitor-kubernetes/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/monitor-kubernetes/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/monitor-with-oss/_index.md |
https://support.influxdata.com/ | Network error: SSL certificate not trusted. Use --insecure if site is trusted (e… |
content/enterprise_influxdb/v1/administration/monitor/monitoring-dashboards/_index.md |
https://support.influxdata.com/ | Error (cached): Error (cached) |
content/enterprise_influxdb/v1/administration/monitor/monitoring-dashboards/_index.md |
https://reddit.com/r/influxdb | Error (cached) |
Showing first 20 of 65 warnings. See the workflow run for full results.
Full details: workflow run summary and artifact. Last updated: 2026-08-24 15:20:13 UTC
Contributor
queryExecutor.queriesFailed, queryExecutor.queriesSlow, and tsm1_engine's six compactionPlanner* fields were removed as "confirmed absent" based on a v1.12.4 test that only wrote one point and ran one query. Re-verified against live InfluxDB v1.13.0 (OSS) and v1.13.0rc5 (Enterprise) with a deliberately failing query and confirmed all 8 fields are present with real, non-zero values.
Contributor
Author
Correction: 8 fields were wrongly removed as "confirmed absent"This PR originally removed That conclusion was wrong. The original test wrote a single point and ran a single query — not enough to trigger a TSM compaction cycle or reach the failed-query code path that populates those fields. Re-verification, done while separately verifying
All 8 fields are restored in the latest commit, with a correction note added next to the existing Everything else in the original verification still holds: the 4 fields flagged |
jstirnaman
commented
Aug 20, 2026
…ents-internal.md Delete lines ported from platform/monitoring docs
jstirnaman
commented
Aug 21, 2026
Contributor
Author
There was a problem hiding this comment.
The follow-up PR migrates this content into the v1 documentation.
jstirnaman
added a commit
that referenced
this pull request
Aug 21, 2026
… sections
- Removed content/platform/monitoring/influxdata-platform/ and
content/platform/monitoring/containers/ (11 files): _internal
measurements, SHOW STATS/DIAGNOSTICS, Kapacitor measurements,
monitoring dashboards, watcher-of-watchers setup, internal-vs-external
monitoring, and Kubernetes monitoring.
- Added 8 shared source files under
content/shared/influxdb-v1/administration/monitor/ and 16 per-edition
stub pages (content/influxdb/v1/administration/monitor/ and
content/enterprise_influxdb/v1/administration/monitor/) for the same
topics, plus a new performance-metrics-commands.md page carrying the
"Useful performance metrics commands" and `/metrics` endpoint sections
that used to live on server_monitoring.md.
- Deleted content/influxdb/v1/administration/server_monitoring.md; its
menu slot ("Monitor InfluxDB", weight 80, parent Administration) is now
a new hub page at content/influxdb/v1/administration/monitor/_index.md,
matching the hub that already existed on the Enterprise v1 side.
- Restructured both v1 OSS and v1 Enterprise monitor hubs into two
sections: "Monitor InfluxDB itself" (internal metrics, SHOW STATS/
DIAGNOSTICS, logs, diagnostics) vs. "Monitor other systems with
InfluxDB" (the TICK-stack pages) -- the two were previously flat and
undifferentiated.
- Deleted content/platform/monitoring/_index.md; its content folds into
a new "Monitoring" section on content/platform/_index.md, using the
same two-way split, linking out per edition to whatever monitor page
already exists (InfluxDB 3 Core/Enterprise query-system-data, Cloud
Dedicated monitor-your-cluster, OSS v2 metrics reference, and the v1
hubs). Editions with no such page yet (Cloud Serverless, Clustered,
Cloud early-access, Cloud TSM) are omitted rather than stubbed.
- Cross-links between the new shared pages use the `/product/version/`
placeholder and `{{% show-in %}}` shortcode so each edition gets
edition-correct link targets (e.g. the disable-`_internal` config
link points at `/influxdb/v1/administration/config/` on OSS and
`/enterprise_influxdb/v1/administration/configure/config-data-nodes/`
on Enterprise).
- Updated every inbound link to the removed pages across the
repo (chronograf, telegraf is generated and covered by aliases,
release notes, troubleshooting, prometheus support pages, query
language specs).
- Fixed two pre-existing typos surfaced by Vale while touching these
files ("your your", "caused caused").
- Every removed page keeps its old URL(s) as a Hugo alias on its new
destination.
Per review decision on #7557: content/platform/ should only carry an
overview with links to individual editions, not product-specific
how-to content. The old platform/monitoring/ tree mixed two different
concepts -- monitoring InfluxDB itself vs. using InfluxDB/the TICK
stack to monitor other systems -- without distinguishing them, which
made the hub pages hard to navigate and duplicated content that
belongs with each edition's own admin docs.
No content is lost; it moved to content/shared and per-edition pages
where it's discoverable from each edition's own admin/administration
navigation instead of a separate platform-level tree. All old URLs
redirect via Hugo aliases.
Out of scope, tracked as follow-up: establishing a real /admin/monitor/
URL convention for InfluxDB 3 editions (today it's scattered
query-system-data/monitor-your-cluster pages at ordinary admin/ paths);
filling the monitoring gap for Cloud Serverless, Clustered, Cloud
(early access), and InfluxDB Cloud (TSM), none of which have any
"monitor the instance itself" documentation today.
- `npx hugo --quiet --gc` builds clean.
- `link-checker map` + `link-checker check` on every changed/added file
and the pages that reference them -- no internal link or fragment
errors (only pre-existing external-URL noise and expected
not-yet-deployed self-links).
- `.ci/vale/vale.sh --config=.vale.ini --minAlertLevel=error` passes on
every touched file.
- Spot-checked redirects: /platform/monitoring/,
/platform/monitoring/influxdata-platform/,
/influxdb/v1/administration/server_monitoring/, and each individual
old tool URL all resolve to their new destination.
- Confirmed menus auto-trim: no leftover "InfluxData Platform" or
monitor-platform menu references anywhere after deleting the pages
(this repo's menus are built purely from surviving pages' frontmatter,
no separate menu data file to prune).
Builds on #7685 (the _internal field corrections that PR verifies live
against InfluxDB v1.12.4); based on that branch so the diff here is
pure reorganization, not content changes.
jstirnaman
added a commit
that referenced
this pull request
Aug 21, 2026
Per review decision on PR #7685: don't hand-write cross-edition monitoring links in page prose (a site-wide banner already flags older editions); instead use alt_links frontmatter so the product/version dropdown routes to the right page when a reader switches products. - measurements-internal.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data, Cloud Dedicated monitor-your-cluster, and OSS v2's metrics reference. - show-stats.md and show-diagnostics.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data -- the closest equivalent, since v3 replaces these InfluxQL statements with system tables. OSS v1 <-> Enterprise v1 needs no alt_links: both editions publish these pages at the identical relative path, so the product selector's default same-path fallback already resolves correctly between them. The other 5 shared pages (Kapacitor measurements, monitoring dashboards, watcher-of-watchers setup, internal-vs-external, monitor Kubernetes) have no equivalent anywhere outside the v1 TICK stack, so no alt_links were added there -- the selector's default of falling back to the target product's root is the correct behavior.
The influxdb input plugin link pointed at the plugin's GitHub README. Point it at the plugin's own docs page instead, matching every other Telegraf plugin link in this PR.
jstirnaman
added a commit
that referenced
this pull request
Aug 21, 2026
… sections
- Removed content/platform/monitoring/influxdata-platform/ and
content/platform/monitoring/containers/ (11 files): _internal
measurements, SHOW STATS/DIAGNOSTICS, Kapacitor measurements,
monitoring dashboards, watcher-of-watchers setup, internal-vs-external
monitoring, and Kubernetes monitoring.
- Added 8 shared source files under
content/shared/influxdb-v1/administration/monitor/ and 16 per-edition
stub pages (content/influxdb/v1/administration/monitor/ and
content/enterprise_influxdb/v1/administration/monitor/) for the same
topics, plus a new performance-metrics-commands.md page carrying the
"Useful performance metrics commands" and `/metrics` endpoint sections
that used to live on server_monitoring.md.
- Deleted content/influxdb/v1/administration/server_monitoring.md; its
menu slot ("Monitor InfluxDB", weight 80, parent Administration) is now
a new hub page at content/influxdb/v1/administration/monitor/_index.md,
matching the hub that already existed on the Enterprise v1 side.
- Restructured both v1 OSS and v1 Enterprise monitor hubs into two
sections: "Monitor InfluxDB itself" (internal metrics, SHOW STATS/
DIAGNOSTICS, logs, diagnostics) vs. "Monitor other systems with
InfluxDB" (the TICK-stack pages) -- the two were previously flat and
undifferentiated.
- Deleted content/platform/monitoring/_index.md; its content folds into
a new "Monitoring" section on content/platform/_index.md, using the
same two-way split, linking out per edition to whatever monitor page
already exists (InfluxDB 3 Core/Enterprise query-system-data, Cloud
Dedicated monitor-your-cluster, OSS v2 metrics reference, and the v1
hubs). Editions with no such page yet (Cloud Serverless, Clustered,
Cloud early-access, Cloud TSM) are omitted rather than stubbed.
- Cross-links between the new shared pages use the `/product/version/`
placeholder and `{{% show-in %}}` shortcode so each edition gets
edition-correct link targets (e.g. the disable-`_internal` config
link points at `/influxdb/v1/administration/config/` on OSS and
`/enterprise_influxdb/v1/administration/configure/config-data-nodes/`
on Enterprise).
- Updated every inbound link to the removed pages across the
repo (chronograf, telegraf is generated and covered by aliases,
release notes, troubleshooting, prometheus support pages, query
language specs).
- Fixed two pre-existing typos surfaced by Vale while touching these
files ("your your", "caused caused").
- Every removed page keeps its old URL(s) as a Hugo alias on its new
destination.
Per review decision on #7557: content/platform/ should only carry an
overview with links to individual editions, not product-specific
how-to content. The old platform/monitoring/ tree mixed two different
concepts -- monitoring InfluxDB itself vs. using InfluxDB/the TICK
stack to monitor other systems -- without distinguishing them, which
made the hub pages hard to navigate and duplicated content that
belongs with each edition's own admin docs.
No content is lost; it moved to content/shared and per-edition pages
where it's discoverable from each edition's own admin/administration
navigation instead of a separate platform-level tree. All old URLs
redirect via Hugo aliases.
Out of scope, tracked as follow-up: establishing a real /admin/monitor/
URL convention for InfluxDB 3 editions (today it's scattered
query-system-data/monitor-your-cluster pages at ordinary admin/ paths);
filling the monitoring gap for Cloud Serverless, Clustered, Cloud
(early access), and InfluxDB Cloud (TSM), none of which have any
"monitor the instance itself" documentation today.
- `npx hugo --quiet --gc` builds clean.
- `link-checker map` + `link-checker check` on every changed/added file
and the pages that reference them -- no internal link or fragment
errors (only pre-existing external-URL noise and expected
not-yet-deployed self-links).
- `.ci/vale/vale.sh --config=.vale.ini --minAlertLevel=error` passes on
every touched file.
- Spot-checked redirects: /platform/monitoring/,
/platform/monitoring/influxdata-platform/,
/influxdb/v1/administration/server_monitoring/, and each individual
old tool URL all resolve to their new destination.
- Confirmed menus auto-trim: no leftover "InfluxData Platform" or
monitor-platform menu references anywhere after deleting the pages
(this repo's menus are built purely from surviving pages' frontmatter,
no separate menu data file to prune).
Builds on #7685 (the _internal field corrections that PR verifies live
against InfluxDB v1.12.4); based on that branch so the diff here is
pure reorganization, not content changes.
jstirnaman
added a commit
that referenced
this pull request
Aug 21, 2026
Per review decision on PR #7685: don't hand-write cross-edition monitoring links in page prose (a site-wide banner already flags older editions); instead use alt_links frontmatter so the product/version dropdown routes to the right page when a reader switches products. - measurements-internal.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data, Cloud Dedicated monitor-your-cluster, and OSS v2's metrics reference. - show-stats.md and show-diagnostics.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data -- the closest equivalent, since v3 replaces these InfluxQL statements with system tables. OSS v1 <-> Enterprise v1 needs no alt_links: both editions publish these pages at the identical relative path, so the product selector's default same-path fallback already resolves correctly between them. The other 5 shared pages (Kapacitor measurements, monitoring dashboards, watcher-of-watchers setup, internal-vs-external, monitor Kubernetes) have no equivalent anywhere outside the v1 TICK stack, so no alt_links were added there -- the selector's default of falling back to the target product's root is the correct behavior.
jstirnaman
changed the title
… sections
- Removed content/platform/monitoring/influxdata-platform/ and
content/platform/monitoring/containers/ (11 files): _internal
measurements, SHOW STATS/DIAGNOSTICS, Kapacitor measurements,
monitoring dashboards, watcher-of-watchers setup, internal-vs-external
monitoring, and Kubernetes monitoring.
- Added 8 shared source files under
content/shared/influxdb-v1/administration/monitor/ and 16 per-edition
stub pages (content/influxdb/v1/administration/monitor/ and
content/enterprise_influxdb/v1/administration/monitor/) for the same
topics, plus a new performance-metrics-commands.md page carrying the
"Useful performance metrics commands" and `/metrics` endpoint sections
that used to live on server_monitoring.md.
- Deleted content/influxdb/v1/administration/server_monitoring.md; its
menu slot ("Monitor InfluxDB", weight 80, parent Administration) is now
a new hub page at content/influxdb/v1/administration/monitor/_index.md,
matching the hub that already existed on the Enterprise v1 side.
- Restructured both v1 OSS and v1 Enterprise monitor hubs into two
sections: "Monitor InfluxDB itself" (internal metrics, SHOW STATS/
DIAGNOSTICS, logs, diagnostics) vs. "Monitor other systems with
InfluxDB" (the TICK-stack pages) -- the two were previously flat and
undifferentiated.
- Deleted content/platform/monitoring/_index.md; its content folds into
a new "Monitoring" section on content/platform/_index.md, using the
same two-way split, linking out per edition to whatever monitor page
already exists (InfluxDB 3 Core/Enterprise query-system-data, Cloud
Dedicated monitor-your-cluster, OSS v2 metrics reference, and the v1
hubs). Editions with no such page yet (Cloud Serverless, Clustered,
Cloud early-access, Cloud TSM) are omitted rather than stubbed.
- Cross-links between the new shared pages use the `/product/version/`
placeholder and `{{% show-in %}}` shortcode so each edition gets
edition-correct link targets (e.g. the disable-`_internal` config
link points at `/influxdb/v1/administration/config/` on OSS and
`/enterprise_influxdb/v1/administration/configure/config-data-nodes/`
on Enterprise).
- Updated every inbound link to the removed pages across the
repo (chronograf, telegraf is generated and covered by aliases,
release notes, troubleshooting, prometheus support pages, query
language specs).
- Fixed two pre-existing typos surfaced by Vale while touching these
files ("your your", "caused caused").
- Every removed page keeps its old URL(s) as a Hugo alias on its new
destination.
Per review decision on #7557: content/platform/ should only carry an
overview with links to individual editions, not product-specific
how-to content. The old platform/monitoring/ tree mixed two different
concepts -- monitoring InfluxDB itself vs. using InfluxDB/the TICK
stack to monitor other systems -- without distinguishing them, which
made the hub pages hard to navigate and duplicated content that
belongs with each edition's own admin docs.
No content is lost; it moved to content/shared and per-edition pages
where it's discoverable from each edition's own admin/administration
navigation instead of a separate platform-level tree. All old URLs
redirect via Hugo aliases.
Out of scope, tracked as follow-up: establishing a real /admin/monitor/
URL convention for InfluxDB 3 editions (today it's scattered
query-system-data/monitor-your-cluster pages at ordinary admin/ paths);
filling the monitoring gap for Cloud Serverless, Clustered, Cloud
(early access), and InfluxDB Cloud (TSM), none of which have any
"monitor the instance itself" documentation today.
- `npx hugo --quiet --gc` builds clean.
- `link-checker map` + `link-checker check` on every changed/added file
and the pages that reference them -- no internal link or fragment
errors (only pre-existing external-URL noise and expected
not-yet-deployed self-links).
- `.ci/vale/vale.sh --config=.vale.ini --minAlertLevel=error` passes on
every touched file.
- Spot-checked redirects: /platform/monitoring/,
/platform/monitoring/influxdata-platform/,
/influxdb/v1/administration/server_monitoring/, and each individual
old tool URL all resolve to their new destination.
- Confirmed menus auto-trim: no leftover "InfluxData Platform" or
monitor-platform menu references anywhere after deleting the pages
(this repo's menus are built purely from surviving pages' frontmatter,
no separate menu data file to prune).
Builds on #7685 (the _internal field corrections that PR verifies live
against InfluxDB v1.12.4); based on that branch so the diff here is
pure reorganization, not content changes.
Per review decision on PR #7685: don't hand-write cross-edition monitoring links in page prose (a site-wide banner already flags older editions); instead use alt_links frontmatter so the product/version dropdown routes to the right page when a reader switches products. - measurements-internal.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data, Cloud Dedicated monitor-your-cluster, and OSS v2's metrics reference. - show-stats.md and show-diagnostics.md (OSS v1 + Enterprise v1): alt_links to InfluxDB 3 Core/Enterprise query-system-data -- the closest equivalent, since v3 replaces these InfluxQL statements with system tables. OSS v1 <-> Enterprise v1 needs no alt_links: both editions publish these pages at the identical relative path, so the product selector's default same-path fallback already resolves correctly between them. The other 5 shared pages (Kapacitor measurements, monitoring dashboards, watcher-of-watchers setup, internal-vs-external, monitor Kubernetes) have no equivalent anywhere outside the v1 TICK stack, so no alt_links were added there -- the selector's default of falling back to the target product's root is the correct behavior.
…mmar
## What changed
- Repointed every Telegraf plugin link in these files from the plugin's
GitHub README to its docs page under /telegraf/v1/input-plugins/:
influxdb, mem, cpu, system, kapacitor, docker, kubernetes, prometheus.
- Fixed a sentence fragment in performance-metrics-commands.md ("Must
have the `monitor` service enabled" had no subject).
- Fixed a comma splice and a doubled verb ("need require") in
monitor-kubernetes.md.
- Fixed a stray `]` and a dropped article in show-stats.md's
`SHOW STATS FOR <component>` sentence.
- Fixed "will be able detect" (missing "to") and a stray capital in
"Monitoring your TICK stack" in internal-vs-external.md.
- Fixed a typo ("watcher of watcher" -> "watcher of watchers") and a
dropped letter ("infuxdb_" -> "influxdb_") in measurements-internal.md.
- Fixed "its" -> "it's" and "a InfluxDB" -> "an InfluxDB" in
monitor-with-oss.md, plus a copy-paste error that said data flows
"from InfluxDB OSS instances to your InfluxDB OSS account" -- every
other section of that page describes Enterprise instances reporting
to an OSS account, so this one was wrong, not just awkward.
- Reworded the `/metrics` endpoint sentence in both
supported_protocols/prometheus.md files (OSS and Enterprise) to drop
passive voice and the redundant "metrics ... metrics format".
## Why
Two review passes: don't send readers to a GitHub README when we have
our own docs page for the same plugin, and general grammar cleanup
requested on the PR.
## What changed
- Re-categorized 4 pages from "Monitor other systems with InfluxDB" to
"Monitor InfluxDB itself" across both hub pages and
content/platform/_index.md: monitoring-dashboards.md (visualizes
InfluxDB/Enterprise/Kapacitor's own metrics), internal-vs-external.md
and external-monitor-setup.md (strategies/topology for monitoring the
TICK stack itself, not other systems), and kapacitor-measurements.md
(Kapacitor's own internal metrics, the Kapacitor analog of
`_internal`). Only monitor-kubernetes.md actually fit "monitor other
systems" -- it's now the section's sole entry.
- Fixed two dead self-referencing anchors in measurements-internal.md
and internal-vs-external.md: Hugo's goldmark slugger preserves the
underscore in headings containing `_internal`, so the real ids are
`disable-the-_internal-database-in-production-clusters` and
`influxdb-_internal-measurements-and-fields`, not the underscore-less
versions the links used.
- Fixed `#retry-concurrency-20`, a cross-page anchor to
config-data-nodes.md that never existed -- the real anchor is
`#retry-concurrency`.
- Added a missing TOC entry for `httpd.valuesWrittenOK` (present in the
body, absent from the field index) and reordered the httpd TOC block
to match body order.
- Joined a 3-line sentence fragment under `cluster.openConnections`
into one line (semantic line feeds).
- Fixed two truncated-word typos ("...destination node fr") in the
`hh_database` and `hh_processor` sections.
- Fixed "Kapacaitor"/"Kapactor" typos in external-monitor-setup.md.
- Repointed platform/troubleshoot/disk-usage.md and oom-loops.md from a
hardcoded OSS-v1-only monitoring-dashboards link to the neutral
`/platform/#monitoring` anchor, matching troubleshoot/_index.md's
existing pattern -- these are edition-agnostic pages.
## Why
Final review pass from the influxdb1-tech-writer agent against both
stacked PRs. All findings verified against the actual rendered HTML
(id= attributes) before fixing, not just the link-checker report, since
link-checker's directory-URL fragment resolution has a known false-
positive bug (tracked in influxdata/docs-tooling#278).
Co-authored-by: Jason Stirnaman <stirnamanj@gmail.com>
Contributor
Release version check
💡 Badge new features with the versionDocumenting a new feature? Add a version badge in the page frontmatter — the
For inline version text, use |
github-actions
Bot
added
product:chronograf
davidby-influx
left a comment
davidby-influx
left a comment
There was a problem hiding this comment.
Should this have userquerybytes and userwritebytes?
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #7557
What changed
queryExecutor.queriesFailed,queryExecutor.queriesSlow, andtsm1_engine's sixcompactionPlanner*fields as "confirmed absent". I re-verified against live InfluxDB v1.13.0 (OSS) and v1.13.0rc5 (Enterprise) with a deliberately failing query; all 8 fields are present with real, non-zero values.shard.indexType: it's a tag (confirmed viaSHOW STATS' tag set), not a field. Moved it into a tags note on the measurement intro instead of a fake field heading.write.pointReqLocal; it's present on OSS too.cluster.openConnections,httpd.fluxQueryRespBytes.fluxQueryDurationNs->fluxQueryReqDurationNs).#pointsWrittenOK/#valuesWrittenOKdon't match Hugo's lowercased, deduped heading IDs (thehttpdandwritemeasurements each have a same-named field, so the real targets are#pointswrittenok-1and#valueswrittenok-1).<!-- VERIFY (live instance) -->comment for the four fields and one measurement that remain unconfirmed, tracked in Verify possibly-retired InfluxDB_internalfields: httpd.valuesWrittenOK, write.pointsWrittenOK/subWriteDrop/valuesWrittenOK, hh_database #7684, instead of shipping speculative status to readers.Why
httpd.valuesWrittenOK,write.pointsWrittenOK,write.subWriteDrop,write.valuesWrittenOKare still documented by the Telegrafinfluxdbinput plugin, so rather than deleting them outright they're flagged as unconfirmed pending engineering input (#7684) — they may be limited to older 1.x releases. That part still holds.The original PR also removed 8 other fields as "confirmed absent," but that conclusion didn't survive a second look — see the correction above.
Impact
Readers get a corrected field/tag list; the four still-uncertain fields no longer read as authoritative in the rendered page. The 8 restored fields (
queryExecutor.queriesFailed/queriesSlow,tsm1_engine'scompactionPlanner*set) are real and now documented again.Verification
Original: live-tested against InfluxDB OSS v1.12.4 (port 8087) and InfluxDB Enterprise v1.12.4 (ports 8086/8091), both freshly started for this check:
SHOW MEASUREMENTS/SHOW FIELD KEYSagainst_internalon both instances, before and after writing line protocol data and running queries (including a deliberately failing one, to testqueryExecutor.queriesFailed).SHOW STATS(instant snapshot, unaffected by the ~10s_internalwrite interval) cross-checked against the persisted_internaldatabase after waiting out that interval — same result both ways.hh_databasecouldn't be tested: reproducing it needs an active hinted handoff failure between cluster nodes, not reproducible with a single-node Enterprise setup.npx hugo --quiet --gc,.ci/vale/vale.sh, andlink-checkerall pass on the changed file.Follow-up (restoring the 8 fields): re-tested against InfluxDB OSS v1.13.0 GA (port 8087) and InfluxDB Enterprise v1.13.0rc5 (ports 8086/8091):
SELECT * FROM nonexistent_measurement_xyz) on both instances, then waited past the_internalwrite interval.SHOW FIELD KEYSagainst_internalon both instances showsqueryExecutor.queriesFailed,queryExecutor.queriesSlow, and all sixtsm1_engine.compactionPlanner*fields present.SELECT last(...)on those fields confirms real, non-zero values (for examplequeriesFailed = 2, matching the one failing query;compactionPlannerPlan = 869), not just declared-but-empty fields.Follow-up
Base for a second PR (
platform->monitormigration) that moves this content into per-edition shared docs and removescontent/platform/monitoring/.