A mis-tagged funnel step corrupted last month's permit metrics. Raw logs still exist. What resilience action should the analyzer support?
Select an answer to reveal the explanation.
Short Explanation
Messed-up tags are fixable when you kept the raw tape. Correct the tracking, replay the logs, and rebuild the funnel like rewashing a load that used the wrong soap. Backfill is how analyzers recover instead of living with a bad month forever.
Full Explanation
When mis-tagged funnel steps corrupt last month’s permit metrics, the analyzer should support reprocessing or backfill from raw logs after fixing the tagging so corrected metrics can be rebuilt. Raw retention plus replay is resilience for civic analytics mistakes. Cap. 10 maturity includes recovery, not only pretty dashboards for municipal program owners.
Leaving wrong funnel numbers permanent because raw logs must never be replayed abandons correctable error. Deleting the lake and abandoning historical comparison after any tagging mistake is disproportionate and destroys institutional memory. Fixing only dashboard colors without correcting underlying event processing leaves decisions on bad facts about permit performance.
Exam caveat: backfills need versioned metric outputs or as-of dates so consumers know when numbers were restated; silent in-place edits erode trust. Operational check: fix a deliberate bad tag in staging, replay the raw window, publish a restated funnel with a revision note, and confirm downstream dashboards show the corrected month. Document the verified decision in the municipal architecture runbook so future on-call and audit staff inherit the same standard. Document the verified decision in the municipal architecture runbook so future on-call and audit staff inherit the same standard.