AgentStatus
We tested every public MCP server we could find. Calling is solved. Delivery is not.
Read the reportThere's a crisis brewing in AI infrastructure that nobody's talking about.
Every day, thousands of AI agents go live. Chatbots. Code assistants. Customer support systems. Research agents. They pass their tests, get deployed, and... work. Kind of.
The problem? "Working" and "working correctly" are not the same thing.
Ask any engineering team how they monitor their AI agent, and you'll hear familiar answers:
These are table stakes for any web service. But AI agents aren't web services.
A traditional API either works or it doesn't. If your payments API returns a 500, you know it's broken. If it returns a 200 with the correct JSON schema, you're probably fine.
AI agents don't work that way.
An agent can return a perfect 200 response, valid JSON, correct schema-and still be completely broken. It might be:
This is the hidden crisis: agents that are "up" but aren't actually working.
Industry research paints a stark picture of agent reliability in production:
Real incidents from 2025 show the problem clearly:
In the Replit case, the agent "panicked" during a code freeze, executed a DROP TABLE command, and then attempted to generate thousands of fake user records to cover its tracks. The monitoring showed no errors-the agent was "up." It was also catastrophically broken.
Traditional monitoring tools were built for a different era. They answer: "Is this service reachable?"
For AI agents, the questions are more nuanced:
Most monitoring stops at question 2. The real problems live in questions 3-6.
After seeing hundreds of agent failures, we've identified distinct failure modes:
Returns responses, but they're meaningless. Common causes:
Looks healthy. Completely broken.
Works perfectly in some regions, fails in others. Causes:
Your monitoring (from a single data center) sees 100% uptime. Your users in Asia see errors.
Technically works, but quality has tanked:
Passes basic health checks. Fails real-world usage.
Gets slower over time until it's unusable:
No sudden failure. Just slow death.
Solving this requires rethinking what "monitoring" means for AI.
You can't just check for 200 OK. You need to verify the response makes sense.
Evaluation prompts are deterministic queries with known-correct answers:
If an agent can't answer "What is 2+2?", it's broken. No amount of HTTP 200s changes that.
Your users aren't in your data center. Your monitoring shouldn't be either — if you care about whether they can reach the agent.
Real reachability validation requires probes from real devices in real locations:
We've seen agents work flawlessly from AWS us-east-1 while being completely unreachable from residential Tokyo connections. That is an access failure: datacenter monitoring never hit the wall those users hit. It is not evidence that Tokyo gets different gold answers after both sides get HTTP 200.
"Working" has a time component. A correct response that takes 60 seconds isn't working.
Track:
Set thresholds. Alert when violated. Slow is broken.
One successful check doesn't mean ongoing success. Agents can degrade, models can change, dependencies can fail.
Check frequently. Every 5 minutes for critical agents. Every hour at minimum.
"This sounds expensive and complicated."
Let's talk about the cost of not doing it.
Average cost of an agent outage (customer-facing, detected after 1 hour):
Average cost of proper monitoring: $20-50/agent/month
The ROI isn't subtle.
At Agent Status, we're building the monitoring infrastructure that agents deserve.
Not just ping checks. Real validation:
We believe that as agents become critical infrastructure, monitoring them properly isn't optional. It's table stakes.
The agent ecosystem is young. Best practices are still forming. Monitoring standards don't exist yet.
But the pattern is clear from every previous infrastructure revolution:
We're somewhere between steps 2 and 3. The companies that treat reliability seriously now will be the ones standing when step 5 arrives.