Arc 26.03.2 — Critical Compaction Fix

We shipped 26.03.2 last night. This is a critical bug fix release and all users should update immediately.
What Happened
When a partition accumulated more than 30 files, the compaction system split them into sequential batches of 30 files each. Each batch generated its output filename using second-precision timestamps — YYYYMMDD_HHMMSS.
When two or more batches completed within the same second, they produced identical filenames. Each subsequent batch silently overwrote the previous batch's compacted file. Depending on how many batches were affected, up to 84% of ingested data in that partition could be lost.
Who Was Affected
Any measurement with more than 30 files per hour partition. In practice, this affects:
- High-frequency ingestion pipelines — ingestion interval below 1,000ms
- Telegraf configurations with multiple field groups
- Any workload where
max_buffer_age_msis configured below 1,000ms
If you were seeing unexplained data gaps in Grafana dashboards — this is the cause.
The Fix
We added nanosecond precision to compacted output filenames, guaranteeing uniqueness across all sequential batches regardless of how fast they complete. We also updated extractNewestFileTime to parse both old and new filename formats — existing data is unaffected.
26.03.2 is a drop-in replacement for 26.03.1. No configuration changes required.
How to Update
docker pull ghcr.io/basekick-labs/arc:latestFor binary installations, download 26.03.2 from the https://github.com/Basekick-Labs/arc/releases.
A Note on How This Was Found
This bug was found by a real user running Arc against production plant data. They reported unexplained data gaps in their Grafana dashboards within hours of setting up Arc. That report led directly to this fix.
This is exactly why early testing and honest feedback matter. Thank you.
What's Next
Beyond this fix, the same investigation surfaced several additional improvements to the compaction and ingestion layers. These will ship in 26.04.1, currently planned for next week.
If you have questions or are seeing unexpected behavior after updating, find us on Discord or open an issue on https://github.com/Basekick-Labs/arc/issues.
Ready to handle billion-record workloads?
Deploy Arc in minutes. Own your data in Parquet. Use for analytics, observability, AI, IoT, or data warehousing.
