Inky - Email Processing Delay Due to Regional Traffic Imbalance - 2026-08-19 

Summary 

Between 2026-08-19 06:00 UTC and 2026-08-19 06:30 UTC, some Inky customers experienced delays in email processing and delivery. During this period, a disproportionate amount of inbound email traffic was directed to a single processing region, resulting in increased load and slower message analysis. A small number of messages were temporarily deferred, causing email delivery delays of several minutes before normal processing resumed.  

Root Cause 

Inky uses DNS-based traffic distribution to balance email processing across multiple U.S. regions. During the incident, a third-party mail provider began directing a significantly larger portion of traffic to a single Inky region despite normal DNS weighting remaining in place. The resulting increase in load temporarily reduced processing performance within that region and caused a small number of messages to be deferred and retried automatically. Traffic distribution returned to normal without intervention, and processing performance recovered shortly thereafter. ,  

Incident Timeline 

• Identified: 2026-08-19 06:00 UTC - Traffic distribution became heavily concentrated in a single Inky processing region.  

• Customer Impact Observed: 2026-08-19 06:10 UTC - Increased regional load resulted in slower processing and delayed delivery for a small number of messages.  

• Resolved: 2026-08-19 06:30 UTC - Traffic distribution returned to normal and processing performance was restored.  

Preventative Measures 

To reduce the likelihood and impact of similar incidents in the future, we are implementing the following improvements: 

Enhancements to Infrastructure Resiliency 

• Reduced DNS Time-To-Live (TTL) values associated with weighted DNS records from 15 minutes to 3 minutes. This change allows external DNS caches to refresh routing information more quickly when traffic distribution becomes imbalanced.  

Enhancements to Monitoring and Alerting 

• Evaluate additional monitoring capabilities focused on regional traffic distribution and processing performance to provide earlier visibility into short-duration traffic imbalances that may not currently trigger alerts. This aligns with the finding that the event duration and impact level were insufficient to activate existing alerting mechanisms.  

Enhancements to Traffic Management 

• Continue reviewing traffic balancing and regional capacity strategies to further reduce the likelihood that concentrated traffic loads can affect processing performance in any single region. This effort is informed by the incident analysis showing that regional traffic concentration was a key contributing factor.