How CustomInk Cut CloudWatch Costs by 85% while improving signal

Every observability bill eventually gets too big to ignore. The usual fix is to sample or drop telemetry data.

CustomInk cut their CloudWatch costs 85% without disrupting operations. No dev work changed, dashboards kept populating, and their security team was able to continue tracing individual transactions across logs.

In this fireside chat, Will, DevOps Manager at CustomInk, and Vit - Devops Engineer, sits down with Grepr's CEO, Jad Naous, and talks through it honestly: what was driving the bill, why the usual fixes were off the table, who he had to win over, how the rollout really went, and what caught him off guard.

What You Will Learn:

• Keep the signal that matters. How CustomInk’s security team maintained the ability to conduct transaction-level investigations using Grepr-preserved unique Request IDs inside log summaries.
• Cut noise, don’t sample. How CustomInk reduced log noise by 85%, automatically, using Grepr's signal processing engine that condenses repetitive logs into summaries.
• Two destinations, one pipeline. How Grepr forwards summaries to CloudWatch while writing full-fidelity (raw) logs to CustomInk's own S3 bucket, cutting costs by 85% with no changes to developer work flows.