Backup retry storms: How you can improve backup reliability

Retries are a diagnostic signal of policy health rather than an effective treatment: Acronis H2 2025 telemetry shows policies with many retries typically suffer persistent customer-side failures (quota exhaustion, VSS writer failures, connectivity issues) that retries do not fix. Uncapped retries create “retry storms” that waste CPU, I/O and network resources, degrade performance for users and neighboring tenants, and should be limited with meaningful delays and alerts. #Acronis #AzureBackup

Keypoints

Telemetry from Acronis H2 2025 shows policies with 10+ daily runs produced ~4.7 million errors in October 2025, with always_full policies having the highest per-job error rate (10.01%).
The top failure causes during business hours were SpaceQuotaReachedHard (15.9%), MemoErrorVssWriterFail (11.48%), and ConnectToPlatformFailedNetwork (9.11%), all persistent issues not resolved by retries.
Industry guidance (Google Cloud, Microsoft Azure, AWS) treats retries as a controlled response to transient faults and emphasizes limits, observability, and escalation rather than unlimited retries.
Retries consume scheduler capacity and server resources (VSS snapshots, disk reads, network uploads, agent metadata writes) and can produce heavy invisible infrastructure waste despite a single “failed” status in the console.
Statistics show selection bias: higher run-count buckets reflect failing policies (more attempts occur because failures happened), so comparing error rates across run-count buckets does not prove retries are harmful.
Practical recommendations: cap retries by backup type (e.g., always_incremental: max 2 retries; always_full: max 1 retry), set delays, alert when a policy exceeds four runs/day, and investigate recurring failures instead of increasing retries.

MITRE Techniques

[None ] No MITRE ATT&CK techniques are mentioned in the article.

Indicators of Compromise

[Domain ] Reference and documentation sources cited – cloud.google.com, docs.aws.amazon.com (and learn.microsoft.com referenced).
[Error/Event name ] Acronis failure codes and diagnostics – SpaceQuotaReachedHard, MemoErrorVssWriterFail, and other error identifiers like ConnectToPlatformFailedNetwork.
[Policy name ] Backup policy identifiers used in analysis and recommendations – always_full, always_incremental (and weekly_full_daily_inc).