Example Workflows¶
Real scenarios showing how to use OpenTrace with your AI assistant.
Investigating a Production Incident¶
You're paged about checkout failures. Here's how to investigate:
1. Understand what's involved
"What services are involved in the checkout flow?"
The AI maps out the services: checkout-api → inventory-service → payment-processor → order-service
2. Check dependencies
"What does checkout-api depend on?"
You see it depends on auth-service, inventory-service, payment-processor, and redis-cache.
3. Trace the failure path
"If inventory-service is slow, what would be affected?"
The AI shows you the blast radius - checkout, order confirmation, and inventory sync would all be impacted.
4. Find the connection
"How does checkout-api connect to the database?"
You see the path goes through order-service which connects to postgres-main.
Understanding a Service Before Making Changes¶
You need to modify the notification service. Start by understanding its role:
1. Get the overview
"Tell me about the notification-service"
2. Find what depends on it
"What services call notification-service?"
You discover 12 services send notifications through it - more than you expected.
3. Check what it needs
"What does notification-service depend on?"
It depends on email-gateway, sms-provider, push-notification-service, and user-preferences-api.
4. Assess the risk
"How critical is notification-service to the system?"
The AI analyzes its connectivity and tells you it's a high-traffic service with many consumers.
Onboarding to a New Codebase¶
You just joined the team. Get up to speed quickly:
1. See the landscape
"What services are in my system?"
2. Find the core services
"What are the most connected services?"
The AI identifies the key services that everything depends on.
3. Understand groupings
"What services handle payments?"
"What services handle user management?"
4. Trace a user journey
"How does a request flow from the mobile app to the order database?"
Planning a Service Deprecation¶
You need to retire the legacy-auth service:
1. Find all consumers
"What services depend on legacy-auth?"
You get a list of everything that needs to be migrated.
2. Check for indirect dependencies
"Are there services that indirectly depend on legacy-auth?"
The AI traces deeper connections you might have missed.
3. Understand the migration scope
"What's the difference between what calls legacy-auth vs new-auth-service?"
4. Verify nothing is missed
"Is there anything else connected to legacy-auth?"
Debugging Performance Issues¶
The API is slow. Figure out why:
1. Map the request path
"Show me the path from api-gateway to the response"
2. Identify bottlenecks
"What databases does the user-profile endpoint touch?"
3. Find shared dependencies
"What do api-gateway, user-service, and order-service all depend on?"
You discover they all share a connection to a cache that might be the bottleneck.
4. Check for cascading issues
"If redis-cache is slow, what services would be affected?"