Maintenance & Deployment Practices

Overview

IMO Health is committed to delivering reliable, high-availability access to our clinical terminology and normalization APIs. Behind every scheduled maintenance window is a disciplined model for safe production change — built on backward/forward compatibility, progressive exposure, automated guardrails, fast rollback, and a disciplined human operating model. This page outlines how we plan, communicate, and execute maintenance so that disruption to your integrations stays minimal and recovery stays fast.

Maintenance Windows

We schedule routine maintenance and production deployments within defined change windows aligned to the criticality of each service. Wherever possible, updates are delivered with zero or minimal downtime using the deployment practices described below.

  • Advance notice — Planned maintenance is communicated ahead of time via email to registered notification recipients.
  • Typical windows — Maintenance is generally performed during low-traffic windows on Tuesdays and Thursdays from 6 to 9 PM CT.
  • Expected impact — Most changes are non-disruptive; when a brief interruption is possible, it is noted in the maintenance notice.
  • Emergency maintenance — In rare cases, urgent fixes may be applied outside the standard window, with notification provided as soon as practical.

Our Deployment Practices

IMO Health production deployments follow a disciplined model for safe change: compatibility + progressive exposure + automated guardrails + fast rollback + disciplined human process. These practices apply to both the technical release mechanisms and the human operating model we use to plan, approve, execute, and monitor every production change.

1. Safe-by-Design Architecture

We design and implement deployment-safe engineering patterns, including:

  • Backward and forward compatibility between service versions
  • Feature flags and kill switches to disable risky behavior changes instantly

2. Progressive Delivery

Critical services follow controlled rollout models such as:

  • Canary deployments with automated health gates
  • Blue/green deployments where rapid cutover or rollback is required
  • Staged rollout rings — from internal users, to limited production cohorts, to full release

3. Automated Release Guardrails

We implement and follow standard release gates that verify:

  • Technical health — latency, error rate, saturation, and dependency health
  • Business health — critical workflow success rates
  • Deployment-specific risk signals
  • On-call engineers are paged automatically when regressions are detected

4. Fast, Tested Rollback

Every critical service has a validated rollback hierarchy:

  • Disable via feature flag or kill switch
  • Route traffic back to the prior healthy version
  • Artifact rollback
  • Infrastructure rollback

5. Human Process Discipline

Our standard release operating model for high-risk changes includes:

  • Deployment risk classification
  • Named release roles
  • Deployment checklists
  • Change windows aligned with service criticality

What This Means for Developers

To make the most of IMO Health's reliability model, we recommend the following for integrations that depend on real-time responses:

  • Implement retry logic with exponential backoff to smoothly handle brief interruptions
  • Subscribe to maintenance notifications via a ticket at support.imohealth.com
  • Review our downtime and service-continuity guidance when designing critical workflows

Notifications & Support

Stay informed and reach us: