Hetzner Online's object storage in data center HEL1 has been partially degraded since 2026-09-09 12:00 UTC, affecting a subset of buckets with elevated response times, stalled requests and occasional timeouts. Engineers identified intermittent stalls in low-level commit operations that can last several seconds and shift across different parts of the cluster, complicating diagnosis. A temporary workaround reduced direct bucket impact, but performance has not fully stabilized: requests continue to encounter delays and some operations still time out.
Remediation efforts include continued investigation and monitoring, containment via the workaround, and active background bucket migrations within HEL1 to relieve I/O pressure; these migrations are performed without required customer changes. The cluster was brought back online and remains under close observation, yet updates through 2026-10-08 indicate persistent elevated response times driven by increased Object Storage utilization. Work is ongoing to restore full service performance and determine the root cause.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.