Summary
Status: Mixed (early-stage — insufficient baseline for trend judgment)
The labelling system became operational on 2026-03-31 and has run consistently since, but only ~4 days of history exist. Automation is working and the correction backlog is currently empty, but multiple summary issues are being created per run day and one run on March 31 shows anomalous change counts (likely a batch initialization), making trend comparison unreliable until a stable baseline forms.
Key Metrics
| Metric |
Value |
Notes |
| Discussions reviewed (last 7 days) |
~120–130 (estimated) |
5 Label Discussions runs across 2 active days; deduplication needed (see detail below) |
| Label changes applied (last 7 days) |
~52–173 (wide range) |
Inflated by a likely batch-init run (#25: 121 changes for 30 discussions); conservative estimate excludes it |
| Change rate (most recent canonical run) |
50% (20 / 40) |
Run #27, 2026-04-03; prior run on 2026-03-31 was ~65% (26 / 40) |
Label Discussions runs (last 7 days) |
5 (all ✅ success) |
runs #21–#23 on 2026-03-31; #26–#27 on 2026-04-03 |
Labelling Correction Collector runs (last 7 days) |
3 (all ✅ success) |
All on 2026-03-31 |
Labelling Correction Feedback runs (last 7 days) |
9 (1 success, 8 skipped) |
Mostly skipped — no open signals to process |
| Open correction signals |
0 |
Both signals from the first batch are closed |
| Signals created (last 7 days) |
2 |
#27, #29 — both created 2026-03-31, closed 2026-04-03 |
| Signals created (last 30 days) |
2 |
Same two |
| Oldest open correction signal age |
N/A |
No open signals |
⚠️ Precision caveat: The reviewed and changed fields did not parse from several summary issues (#25, #28, #33, #34) through the structured parser — values above were extracted from issue body text. Multiple summary issues per run day (3 on 2026-03-31, 2 on 2026-04-03) complicate aggregation; canonical per-day runs are used where identifiable.
Correction Pressure
Correction pressure is minimal at this stage — only 2 signals were generated in the entire observed window.
- Discussion
#98 (category: Other Feature Feedback, Questions & Ideas): received 2 correction events, with Actions as the latest label applied. This indicates the Actions label may have been applied in a context where the category or label mapping was ambiguous.
- Discussion
#118 (category: Enterprise): received 1 correction event, with bug as the latest label applied.
Neither signal is currently open. Pressure is concentrated in a single parent intake issue (Batch 01, #26), now closed. There is no multi-parent spread yet — the system is too new to show clustering.
No recurring label or category cluster is identifiable from two data points, but Actions in the Q&A category is worth monitoring in future batches.
Open Instruction Debt
The correction backlog is clean. Both signals from the first batch (Batch 01 / #26) were resolved within 3 days of creation. There are no open parent intake issues and no stale signals.
However, this does not yet reflect a stable operational state — the system was bootstrapped on 2026-03-31 and has not yet accumulated enough cycles to reveal recurring instruction gaps. The next 2–4 weeks will be the true indicator of whether instruction debt is accumulating or remaining low.
Recommendations
-
Resolve the duplicate summary issue problem. Two summary issues were created on 2026-04-03 (#33 immediately closed, #34 open) and three on 2026-03-31 (#25, #28, #31). This makes metric aggregation unreliable. Investigate whether the Label Discussions workflow is triggering multiple times per schedule interval and, if so, add a deduplication or idempotency guard.
-
Audit the 121-change batch-init run (#25). Issue #25 (2026-03-31) recorded 121 label changes across only 30 discussions — far above any other run. This is consistent with a bulk initialization of a previously-unlabeled backlog, but it should be confirmed. If it was a one-time batch-init, it should be tagged/excluded from ongoing rate baselines to avoid distorting future trend comparisons.
-
Monitor the Actions label in Q&A-category discussions. Signal #29 (Discussion #98) received 2 correction events on the Actions label in the Other Feature Feedback, Questions & Ideas category. Review the relevant rules in .github/instructions/community-discussion-labeling.md to confirm the criteria for applying Actions are unambiguous, especially for discussions that mention GitHub Actions but are primarily support questions.
-
Establish a stable weekly baseline before drawing trend conclusions. With only 4 days of data and a noisy first day, the current change rate (50%–65%) cannot yet be classified as improving or regressing. Allow 2–3 more weeks of consistent runs before interpreting directional movement.
Recent daily summary issue breakdown
| Issue |
Date |
State |
Reviewed |
Changed |
Notes |
#34 |
2026-04-03 |
open |
40 |
20 |
Canonical April 3 run (run #27) |
#33 |
2026-04-03 |
closed same day |
40 |
20 |
Duplicate of #34 (run #26); identical counts |
#31 |
2026-03-31 |
closed 2026-04-03 |
10 |
6 |
Partial run (run #23) |
#28 |
2026-03-31 |
closed same day |
40 |
26 |
Normal run (run #22); 14 skipped |
#25 |
2026-03-31 |
closed same day |
30 |
121 |
Anomalous run (run #21); likely batch init |
Open correction signal breakdown
No open correction signals at time of report.
Closed signals (last 7 days):
| Issue |
Category |
Latest Label |
Corrections |
Created |
Closed |
#29 |
Other Feature Feedback, Questions & Ideas |
Actions |
2 |
2026-03-31 |
2026-04-03 |
#27 |
Enterprise |
bug |
1 |
2026-03-31 |
2026-04-03 |
Parent intake issues:
| Issue |
State |
Created |
Closed |
#26 — Batch 01 |
closed |
2026-03-31 |
2026-04-03 |
Recent workflow run references
| Workflow |
Run |
Date |
Result |
| Label Discussions |
§27 |
2026-04-03 |
✅ success |
| Label Discussions |
§22 |
2026-03-31 |
✅ success |
| Labelling Correction Collector |
§10 |
2026-03-31 |
✅ success |
| Labelling Correction Feedback |
§17 |
2026-03-31 |
✅ success |
| Label Discussions |
§21 |
2026-03-31 |
✅ success |
References
§27 · §10 · §17
Generated by Labelling Health Report · ● 438.1K · ◷
Summary
Status: Mixed (early-stage — insufficient baseline for trend judgment)
The labelling system became operational on 2026-03-31 and has run consistently since, but only ~4 days of history exist. Automation is working and the correction backlog is currently empty, but multiple summary issues are being created per run day and one run on March 31 shows anomalous change counts (likely a batch initialization), making trend comparison unreliable until a stable baseline forms.
Key Metrics
Label Discussionsruns across 2 active days; deduplication needed (see detail below)#25: 121 changes for 30 discussions); conservative estimate excludes it#27, 2026-04-03; prior run on 2026-03-31 was ~65% (26 / 40)Label Discussionsruns (last 7 days)#21–#23on 2026-03-31;#26–#27on 2026-04-03Labelling Correction Collectorruns (last 7 days)Labelling Correction Feedbackruns (last 7 days)#27,#29— both created 2026-03-31, closed 2026-04-03Correction Pressure
Correction pressure is minimal at this stage — only 2 signals were generated in the entire observed window.
#98(category: Other Feature Feedback, Questions & Ideas): received 2 correction events, withActionsas the latest label applied. This indicates theActionslabel may have been applied in a context where the category or label mapping was ambiguous.#118(category: Enterprise): received 1 correction event, withbugas the latest label applied.Neither signal is currently open. Pressure is concentrated in a single parent intake issue (Batch 01,
#26), now closed. There is no multi-parent spread yet — the system is too new to show clustering.No recurring label or category cluster is identifiable from two data points, but
Actionsin the Q&A category is worth monitoring in future batches.Open Instruction Debt
The correction backlog is clean. Both signals from the first batch (Batch 01 /
#26) were resolved within 3 days of creation. There are no open parent intake issues and no stale signals.However, this does not yet reflect a stable operational state — the system was bootstrapped on 2026-03-31 and has not yet accumulated enough cycles to reveal recurring instruction gaps. The next 2–4 weeks will be the true indicator of whether instruction debt is accumulating or remaining low.
Recommendations
Resolve the duplicate summary issue problem. Two summary issues were created on 2026-04-03 (
#33immediately closed,#34open) and three on 2026-03-31 (#25,#28,#31). This makes metric aggregation unreliable. Investigate whether theLabel Discussionsworkflow is triggering multiple times per schedule interval and, if so, add a deduplication or idempotency guard.Audit the 121-change batch-init run (
#25). Issue#25(2026-03-31) recorded 121 label changes across only 30 discussions — far above any other run. This is consistent with a bulk initialization of a previously-unlabeled backlog, but it should be confirmed. If it was a one-time batch-init, it should be tagged/excluded from ongoing rate baselines to avoid distorting future trend comparisons.Monitor the
Actionslabel in Q&A-category discussions. Signal#29(Discussion#98) received 2 correction events on theActionslabel in the Other Feature Feedback, Questions & Ideas category. Review the relevant rules in.github/instructions/community-discussion-labeling.mdto confirm the criteria for applyingActionsare unambiguous, especially for discussions that mention GitHub Actions but are primarily support questions.Establish a stable weekly baseline before drawing trend conclusions. With only 4 days of data and a noisy first day, the current change rate (50%–65%) cannot yet be classified as improving or regressing. Allow 2–3 more weeks of consistent runs before interpreting directional movement.
Recent daily summary issue breakdown
#34#27)#33#34(run#26); identical counts#31#23)#28#22); 14 skipped#25#21); likely batch initOpen correction signal breakdown
No open correction signals at time of report.
Closed signals (last 7 days):
#29Actions#27bugParent intake issues:
#26— Batch 01Recent workflow run references
References
§27 · §10 · §17