# Postmortem: Delayed Start Compute Operations

DevFeed: [Postmortem: Delayed Start Compute Operations](<https://devfeed.tech/articles/postmortem-delayed-start-compute-operations-5743.md>)

Original publisher: [Read original article](<https://neon.com/blog/postmortem-delayed-start-compute-operations>)

Author: Em Sharnoff

Published: 2025-05-21T18:44:44Z

Content type: article

Language: en

Sources: [Blog -- Neon Docs](<https://devfeed.tech/sources/blog-neon-docs.md>)

Topics: [incident](<https://devfeed.tech/topics/incident.md>), [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [Database](<https://devfeed.tech/topics/database.md>), [VPC](<https://devfeed.tech/topics/vpc.md>), [Kubernetes](<https://devfeed.tech/topics/kubernetes.md>), [Query (disambiguation)](<https://devfeed.tech/topics/query.md>), [monitor](<https://devfeed.tech/topics/monitor.md>)

Tags: [aws](<https://devfeed.tech/tags/aws.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [databases](<https://devfeed.tech/tags/databases.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [incident](<https://devfeed.tech/tags/incident.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [monitor](<https://devfeed.tech/tags/monitor.md>), [outage](<https://devfeed.tech/tags/outage.md>), [production](<https://devfeed.tech/tags/production.md>), [proxy](<https://devfeed.tech/tags/proxy.md>), [rate-limiting](<https://devfeed.tech/tags/rate-limiting.md>), [recovery](<https://devfeed.tech/tags/recovery.md>), [regression](<https://devfeed.tech/tags/regression.md>), [vpc](<https://devfeed.tech/tags/vpc.md>)

## AI overview

Neon's postmortem explains two AWS/us-east-1 incidents that prevented customers from creating or starting inactive databases for 5.5 hours. An expensive control-plane database query overloaded the database, delayed suspension of inactive databases, drove pod counts beyond planned capacity, and exhausted subnet IP addresses. Neon mitigated the incidents through AWS CNI reconfiguration, control-plane database scaling, rate limiting, traffic redirection, and planned Kubernetes architecture changes.

## Source excerpt

Summary On two occasions in the past week, Neon customers in AWS/us-east-1 were unable to create or start inactive databases, for a total period of 5.5 hours. Customers with running databases were unaffected. The root cause related to our ability to assign IP addresses to new dat...