Milestone M35 - Kubecost for billing, Cloudwatch migration and more

• Kubecost Integration POC: Completed a proof of concept to evaluate the use of Kubecost within our Kubernetes cluster. This is a step toward simplifying and improving accuracy in our metered billing system.

• NeetoCI Platform Expansion: Enhanced NeetoCI to support one-time compute tasks without requiring GitHub integration or predefined workflows. The first use case will be within NeetoPlaydash to execute Playwright tests, with plans to expand access to a broader user base.

• Incident Resolution — 5xx Errors on NeetoDeploy: We encountered intermittent 5xx errors across multiple applications on NeetoDeploy. After investigation, the root cause was traced to our CertManager component, which handles auto-renewal of origin certificates. Since Cloudflare edge certificates were also in place, not all requests failed. The CertManager issue has since been resolved, and the 5xx errors have ceased.

• NeetoDeploy Infrastructure Migration: Key backend components of NeetoDeploy — including the Dyno Manager, Addon Manager, and Slug Compiler — have been successfully migrated from Heroku to NeetoDeploy’s native environment. This marks a major step in making our deployment platform fully self-hosted.

• Log System Migration: Initiated the migration from Loki/S3 to AWS Cloudwatch for log management. The migration of live logs is complete; however, we discovered a limitation with Cloudwatch’s live stream functionality, which supports only two concurrent live streams. We are currently exploring solutions to address this limitation. Historical logs migration is underway.

Stay connected for more updates as we continue refining and scaling our systems.