/teams-app-developer

Infrastructure

Transport

Rating: GREEN for HTTP-to-HTTPS, RED for Socket Mode → HTTPS (firewall environments).

Impact

Slack bots using Socket Mode run behind firewalls with zero inbound ports. Teams requires a public HTTPS endpoint — a fundamental architecture change for firewall-restricted environments.

Mitigation (Slack Socket Mode → Teams)

Dual-Bot Transport

For bots targeting both platforms simultaneously:

Compute (AWS ↔ Azure)

Cold Start Warning

Azure Functions Consumption plan has 5–10 second cold starts that violate the Teams 3-second response timeout. Mitigations:

Storage (AWS ↔ Azure)

Bot State Storage

Mitigation: Implement the Teams IStorage interface with Cosmos DB for bot state. Use serverless pricing for development, provisioned RUs for production.

Secrets & Configuration

Bot Credentials

Production Secret Management

Observability (AWS ↔ Azure)

Bot Health Monitoring

Key metrics to track for both platforms:

Setup

// Application Insights — must be first import
import appInsights from "applicationinsights";
appInsights.setup(process.env.APPLICATIONINSIGHTS_CONNECTION_STRING).start();

// Then import everything else
import { App } from "@microsoft/teams.apps";

Pitfall: Late instrumentation import. applicationinsights must run before http/https are loaded or distributed tracing won't work.

Rate Limiting & Resilience

Rating: GREEN for basic rate limiting, YELLOW for resilience patterns.

Mitigation (Slack → Teams)

Build a RetryPlugin with exponential backoff + jitter:

Effort: 12–16 hrs for a production-grade retry plugin.

Reverse Direction (Teams → Slack)

Use Bolt's built-in retryConfig option:

const app = new App({
  token: process.env.SLACK_BOT_TOKEN!,
  signingSecret: process.env.SLACK_SIGNING_SECRET!,
  // Built-in retry with exponential backoff
});