September 2018 Summaries

Harness achieves zero downtime deployments by employing a strategy that supports multiple active versions of its Manager and Delegates, which ensures compatibility and seamless upgrades without disrupting ongoing customer executions. This approach, similar to but more flexible than Blue-Green deployment, allows for instant rollback and frequent production deployments, with rollbacks utilized twice in the last four weeks. The architecture involves a Manager that orchestrates pipeline and workflow executions, coordinating with Delegates that run on customers' premises, allowing complex workflows without accessing customer resources directly. The system handles challenges such as varying network connectivity and upgrade times for Delegates by ensuring all involved processes are compatible, using version-aware entities and HTTP header-based routing for communication. This method allows Harness to deploy updates daily while maintaining service continuity, with a rollback mechanism that quickly reverts to previous versions if necessary.

Integrating Bugsnag with Harness Continuous Delivery significantly enhances deployment reliability by automatically detecting new errors or exceptions post-deployment, allowing for immediate rollback to stable versions if issues arise. This integration supports canary deployment strategies, ensuring a gradual and safe rollout of new code while Bugsnag provides deep diagnostic data such as stack traces and session parameters, making it a valuable tool for application developers, DevOps engineers, and SREs. Harness utilizes machine learning algorithms to analyze data sets from Bugsnag, automatically flagging anomalies and regressions and eliminating the need for context switching between tools. This setup allows for comprehensive verification of performance, quality, and security in deployment workflows using various tools, including AppDynamics, New Relic, Dynatrace, Datadog, Splunk, and ELK. Configuring Bugsnag in Harness involves setting up a connector with an API URL and Auth token, enabling the addition of Bugsnag verification to any deployment workflow, and supporting phased canary deployments by progressively increasing infrastructure exposure based on error-free performance.

Harness has enhanced its Continuous Delivery platform by adding LDAP integration, which streamlines the synchronization of user authentication and authorization with existing LDAP providers, simplifying security management for organizations. This addition is part of Harness's focus on providing robust security capabilities, complementing its existing support for Role-Based Access Control (RBAC) and SAML, which were introduced in March 2018. The integration process involves configuring LDAP for authentication and authorization, allowing admins to map and sync user groups between Harness and LDAP providers with ease. This integration offers organizations flexibility, as they can choose between using LDAP, native Harness authentication, or their own SAML provider for user authentication.

Harness Continuous Delivery streamlines deployment pipelines for Red Hat OpenShift by replacing manual Jenkins Pipelines with automated workflows, offering advanced deployment strategies such as Blue-Green and Canary through AI-driven Continuous Verification to ensure performance and quality. OpenShift, an enterprise-grade application platform using Kubernetes, serves as an abstraction layer between applications and underlying infrastructure, providing an easy management option for those using Red Hat Enterprise Linux. While OpenShift Pipelines rely heavily on Jenkins and customer-written scripts, Harness offers a more dynamic alternative with templates and supported plugins, eliminating the need for extensive manual scripting. Harness integrates with various Application Performance Monitoring (APM) and Log Analytics tools, allowing for the automatic verification of deployments and facilitating rollback in case of anomalies. The process to set up an OpenShift deployment pipeline with Harness is designed to be efficient, taking only 5-10 minutes, and includes creating services, cloud providers, workflows, pipelines, and triggers.