Sequoia Capital
Menlo Park, California
Financial Services / Venture Capital
Amazon CloudWatch Logs Amazon CloudWatch Metrics & Alarms Amazon SNS Amazon ECS on AWS Fargate AWS IAM Fluent Bit / FireLens Terraform / Infrastructure-as-Code
Sequoia Capital is one of the world’s most recognized venture capital firms, managing a complex AWS infrastructure to support its engineering operations. Despite using Datadog for infrastructure monitoring and log aggregation, Sequoia’s actual usage was limited to basic CPU and memory threshold alerts and log forwarding, a functionality well within the capabilities of AWS-native services. Facing a contract renewal deadline, Sequoia engaged Avahi to migrate its observability stack from Datadog to Amazon CloudWatch across three AWS environments, eliminating unnecessary third-party costs while maintaining full monitoring coverage. The result was a clean, fully IaC-managed migration completed before the renewal deadline, leaving Sequoia’s team with an auditable, maintainable, and cost-effective observability solution they own outright.
Sequoia Capital is a leading global venture capital firm headquartered in Menlo Park, California, with decades of experience backing transformative technology companies across every stage of growth. Their engineering team manages AWS-based infrastructure supporting internal and public-facing operations, running containerized workloads on Amazon ECS on AWS Fargate across multiple environments.
Sequoia Capital’s engineering team had been using Datadog as their observability platform for infrastructure monitoring and log aggregation. In practice, however, their usage was narrow: CPU and memory threshold alerts, and log forwarding. Datadog’s more advanced capabilities (dashboards, APM traces, synthetic uptime monitors, and anomaly detection) went entirely unused.
As Datadog’s annual renewal approached, Sequoia’s leadership recognized a clear misalignment between what they were paying for and what they were actually using. AWS’s own native observability services — Amazon CloudWatch Logs and CloudWatch Alarms — are fully capable of handling the same workload at a significantly lower cost. Continuing to renew the Datadog subscription meant paying a premium for a sophisticated third-party platform to perform work that their existing AWS environment could handle natively.
The challenge was not simply a cost decision. It was an operational one. Sequoia needed to execute the migration carefully enough to ensure zero monitoring gaps during the transition across three separate AWS environments (development, staging, and production), all before a hard mid-November renewal deadline. Any lapse in alerting coverage during cutover could have left their team blind to critical infrastructure issues in production. They also needed to exit cleanly: all Datadog-specific configurations removed from the codebase, and their team fully equipped to own and maintain the new solution independently.
Sequoia Capital’s infrastructure was already fully AWS-native, running containerized workloads on Amazon ECS on Fargate. The logical path forward was to consolidate observability within the same ecosystem rather than continuing to rely on a third-party tool for capabilities AWS provides natively. Amazon CloudWatch delivers centralized log aggregation, metric-based alerting, and notification routing, covering exactly the monitoring functions Sequoia required, at a cost structure that scales with usage rather than a fixed enterprise subscription.
By moving to CloudWatch, Sequoia also gained tighter integration with their existing AWS environment, eliminated a third-party dependency in their operational stack, and ensured their monitoring infrastructure could be managed and reproduced through the same Terraform workflows they already used for everything else.
Avahi is a premier-tier AWS Partner with deep expertise in AWS-native infrastructure, migrations, and Infrastructure-as-Code. Sequoia needed a partner who could move quickly, work directly within their existing Terraform repositories, and execute a zero-gap migration against a fixed deadline, with no room for monitoring failures in production.
Avahi brought a structured migration methodology specifically suited to this type of transition: a one-to-one monitor audit and alarm mapping approach that eliminated guesswork by cataloguing every active Datadog monitor and recreating it as an exact CloudWatch equivalent. This precision-driven process, combined with Avahi’s IaC-first implementation philosophy, gave Sequoia confidence that nothing would be lost in the transition and that every change would be fully auditable and maintainable by their own team after handoff.
Avahi executed the migration in four sequential phases, designed to validate each layer of the observability stack before advancing to the next environment.
Phase 1 – Kickoff and Discovery: Avahi gained access to Sequoia’s AWS accounts, Terraform repositories, and active Datadog monitor configurations. All log sources were inventoried across development, staging, and production, and a structured migration plan was established with clear milestones for each environment cutover.
Phase 2 – Logging Pipeline Migration: Avahi assessed Sequoia’s existing Fluent Bit and FireLens configurations, which were forwarding container logs to Datadog, and redesigned the logging architecture using Amazon CloudWatch Logs. New log groups were established with defined naming conventions and retention policies for all applications and environments, including public-facing website logs. ECS task definitions and IaC templates were updated to redirect all logs to CloudWatch, with completeness verified in development before rolling out to staging and production.
Phase 3 – Monitoring and Alerting Migration: Avahi conducted a full audit of every active Datadog monitor, documenting threshold values, evaluation periods, affected resources, and notification recipients. Each monitor was mapped one-to-one to an equivalent CloudWatch Alarm with identical trigger conditions, eliminating the risk of misconfigured thresholds or coverage gaps. Alarms were implemented via Terraform starting in development, with Amazon SNS topics configured to route alerts to Sequoia’s existing Slack channels and email distribution lists. Alarm functionality was validated by simulating trigger conditions before rollout to staging and production.
Phase 4 – Cutover and Decommissioning: A formal production cutover runbook was developed and executed. To mitigate the risk of a monitoring gap, both systems were run in parallel during a brief validation window in non-production environments before the final production switchover. Once CloudWatch logging and alerting were confirmed 100% operational across all three environments, all Datadog-specific configurations were removed from the codebase.
Sequoia Capital successfully exited its Datadog contract before the renewal deadline, eliminating the recurring subscription cost entirely. The migration delivered equivalent monitoring and alerting coverage using AWS-native services, with no gaps in visibility during or after the transition. Because every change was implemented as code within Sequoia’s existing Terraform repositories, the new observability stack is fully reproducible, auditable, and owned entirely by their engineering team.
The migration framework Avahi developed (monitor audit, one-to-one alarm mapping, IaC-first implementation, and environment-sequenced rollout) is directly reusable, and Sequoia’s team is now fully equipped to scale and maintain their CloudWatch observability stack independently.
Let’s explore your high-impact AI opportunities together in a complimentary session
