In the world of robotics, data is oxygen. But trying to stream all your data from edge to cloud? That's a surefire way to suffocate your bandwidth—and your budget.
What We're Announcing
Today, we're thrilled to release Rolling Buffer and Device Recordings in INSAION.
The paradigm shift: Capture high-fidelity data locally → Sync to cloud only when you need it.
The Data Dilemma: Feast or Famine
Robotics teams face an impossible choice:
Option 1: Stream Everything 📡
The approach: Use 4G/5G to stream all logs and telemetry
The problems:
- ❌ Incredibly expensive
- ❌ Often technically impossible with high-bandwidth sensors (LiDAR, cameras)
Option 2: Stream Almost Nothing 📉
The approach: Downsample aggressively or only send health heartbeats
The problems:
- ✅ Save money
- ❌ Flying blind when robots behave strangely
The Real Problem
Both options compromise your ability to resolve incidents quickly. You either drown in noise or starve for signal.
Enter the Rolling Buffer
Think of it as a "black box" flight recorder that never runs out of tape.
The Rolling Buffer is a configurable, on-device recording engine that sits right on your robot. It smartly overwrites the oldest data once it hits your set limit.
What You Can Do
⚙️ Configure Policies
Set maximum storage (e.g., 50GB) and retention policy:
- FIFO to cycle old data
- HARD stop when full
🔧 Define Pipelines
Create precise recording rules:
- Record
/odometryat 100Hz - Record
/camera/image_rawonly in specific states - Multiple pipelines with include/exclude patterns
🚀 High Fidelity, Zero Latency
Data stored locally means:
- ✅ No uplink bandwidth constraints
- ✅ Raw sensor data capture
- ✅ High-frequency control loops
- ✅ Zero impact on network budget
The Device Recordings: Surgical Sync
Data at the edge is useful. Data in the cloud is actionable for the whole team.
This is where Device Recordings comes in.
From Firehose to Scalpel
When an incident occurs—say, a navigation failure at 2:00 PM—you don't need the whole day's data.
You need the minutes surrounding the event.
Device Recordings visualizes the timeline of data stored on the device. From there, you can:
1. Browse the Timeline: See exactly what data is available on the robot right now.
2. Select & Sync: Choose specific time blocks or recordings to upload to the cloud.
3. Analyze Instantly: Once synced, the data is available for deep inspection, replay, and analysis by your entire engineering team via the INSAION portal.
Slashing MTTR and Truck Rolls
Rolling Buffer + On-Demand Sync = Operational Game-Changer
1. The "What Happened?" Moment 🔍
Before: Customer reports robot is "acting weird" → You guess
Now: Immediately trigger sync of that specific time window
You get:
- High-fidelity context for race conditions
- Sensor glitch diagnosis
- Edge cases that low-frequency monitoring misses
2. Decrease Mean Time To Recovery (MTTR) ⏱️
Before: Debugging starts days after incident
Now: Debugging starts minutes after incident
You skip: The "can you reproduce this?" phase You reach: Root cause directly
3. Fewer Client Visits 🚗
The biggest win: Reduction in unnecessary trips
Before: Send engineer to site → Plug in laptop → Copy log files
Now: Do it from your desk
Roll a truck only when physical repair is truly necessary.
Completing the Observability Picture
These features integrate seamlessly with INSAION's existing live monitoring and AI-agents. You still get your real-time dashboards for health checks and live status. But now, for those complex, heavy-data problems, you have a powerful new tool in your arsenal.
We believe that better data leads to better robots. With Rolling Buffer and Recordings, we're making sure you have the best data, right when you need it most.
Ready to upgrade your fleet's observability? The Rolling Buffer is available now for all supported agent versions. Check the docs to get started.