Workflow · Production

We Spent $47,000 Running AI Agents in Production. Here's What Nobody Tells You About A2A and MCP.

The problem

The team deployed a multi-agent LangChain system to production believing it would run smoothly, but no guardrails existed to detect runaway agent behavior, and two agents entered an infinite loop that went undetected for 11 days.

First attempt

Two agents in the production multi-agent system got stuck in an infinite conversation loop for 11 days, accumulating $47,000 in API costs before the team shut it down.

Workflow diagram · grounded in source
1
Multi-agent system deployment
trigger
“our team deployed what we thought was a simple multi-agent system to production. Four LangChain agents coordinating via A2A to help users research market data.”
2
Agent-to-agent coordination
ai_action
“Four LangChain agents coordinating via A2A to help users research market data”
3
Infinite loop failure
ai_action
“Two agents got stuck in an infinite conversation loop. For 11 days.”
4
System shutdown
output
“Total damage: $47,000 before we finally pulled the plug.”
Reported outcome

The team shut down the system after $47,000 in total API costs, with weekly costs escalating from $127 in week 1 to $18,400 in week 4.

Reported metrics
Total API costs$47,000
Week 1 API costs$127
Week 2 API costs$891
Week 3 API costs$6,240
Show all 6 reported metrics
Total API costs$47,000
Week 1 API costs$127
Week 2 API costs$891
Week 3 API costs$6,240
Week 4 API costs$18,400
Infinite loop duration11 days
Reported stack
LangChainA2AMCP
Source
https://pub.towardsai.net/we-spent-47-000-running-ai-agents-in-production-heres-what-nobody-tells-you-about-a2a-and-mcp-5f845848de33
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The team shut down the system after $47,000 in total API costs, with weekly costs escalating from $127 in week 1 to $18,400 in week 4.

What tools did this team use?

LangChain, A2A, MCP.

What results were reported?

Total API costs: $47,000; Week 1 API costs: $127; Week 2 API costs: $891; Week 3 API costs: $6,240 (source-reported, not independently verified).

What failed first in this deployment?

Two agents in the production multi-agent system got stuck in an infinite conversation loop for 11 days, accumulating $47,000 in API costs before the team shut it down.

How is this workflow AI workflow structured?

Multi-agent system deployment → Agent-to-agent coordination → Infinite loop failure → System shutdown.