An operations engineer gets a complaint: "the device in the warehouse stopped responding." No SSH access, and the warehouse is 200 km away. How do you find the cause in 30 seconds? IoT device diagnostics via a mobile app is the only quick tool. We specialize in developing such solutions and know how to turn raw metrics into a clear picture. Our team has 5+ years of experience and has built diagnostic panels for 30+ IoT projects, including gateway networks with thousands of devices. A properly designed dashboard reduces fault-finding time from hours to minutes. For example, in one project, 40% of MQTT messages were lost due to weak WiFi (RSSI -85 dBm) — the app identified this in a day, not a week of manual analysis. In another project, 20% of devices rebooted due to overheating — we added a temperature alert, cutting downtime by 50%. Over the course of a year, the app helped prevent 15 emergencies, saving 200 engineer-hours and reducing operational costs by 30–40%. Our stack includes SwiftUI and Combine for iOS (SwiftUI IoT expertise), Jetpack Compose for Android (Jetpack Compose IoT), and Flutter for cross-platform solutions (Flutter IoT). We use MQTT for reliable diagnostics as per the MQTT 3.1.1 specification (OASIS), with TLS for secure data transmission and Firebase for the backend.
Which IoT diagnostic metrics should you collect first?
Practice shows that 80% of issues are covered by 10–15 indicators. Below is the mandatory minimum for effective IoT metric monitoring.
| Metric | Type | Why it matters |
|---|---|---|
| CPU load (core-0, core-1) | float 0–100% | Shows utilization, overload |
| RAM used / total | MB | Memory leaks are a common cause of crashes |
| CPU temperature | °C | Overheating = throttling and failures |
| Uptime | seconds | Reboots are an indirect sign of instability |
| Firmware version | string | To verify against the current one |
| WiFi RSSI | dBm | Connection quality, packet loss |
| MQTT reconnect count | int | Network/broker problems |
| Last error | string | Last error in the log |
On the mobile side, we implement an IoT dashboard with gauge indicators, a live temperature chart, and an error list. Often, we add historical trends for the last 6 hours.
How to choose a data transmission protocol?
For device communication, we use MQTT, REST, or CoAP. For robust IoT diagnostics, we prefer MQTT with QoS 2 and TLS.
| Protocol | Latency | Reliability | Complexity |
|---|---|---|---|
| MQTT | Low | High (QoS 2) | Medium |
| REST | High (polling) | Medium | Low |
| CoAP | Low | Medium | Medium |
In most projects, we choose MQTT with TLS — optimal for mobile diagnostics.
How to interpret RSSI?
RSSI is not an abstract number. Here are the thresholds we embed in the app:
String rssiDescription(int dbm) { if (dbm >= -50) return 'Excellent signal'; if (dbm >= -60) return 'Good signal'; if (dbm >= -70) return 'Satisfactory'; if (dbm >= -80) return 'Weak — possible packet loss'; return 'Critically weak'; } RSSI -80 dBm and below is a reliable cause of periodic MQTT message loss. In one project, the client spent six months looking for a "firmware glitch," but the problem was a remote server with -87 dBm. Adding color indicators to the app cut diagnostic time by 3–5 times.
Why is app diagnostics more effective than standard tools?
Typical alternatives are SSH, web interface, or log files. The app wins in three aspects:
- Speed — opening the app is faster than connecting to VPN and SSH. 3–5 times faster.
- Context — the app itself highlights anomalies (color, thresholds). No need to manually analyze numbers.
- Push notifications — an alert arrives for critical events even when the screen is off (using APNs and FCM push notifications).
We combine data from multiple devices and build aggregated dashboards — e.g., "temperature across all gateways over the last hour." Thanks to real-time diagnostics, our clients reduce downtime by 50% and cut operational costs by 30–40%.
How do we do it?
The typical project process:
- Analysis — we determine which metrics are available on the device (from documentation or test access). We choose the protocol (MQTT, REST, CoAP).
- Design — we sketch the screens, agree on thresholds and visualization. For iOS we use SwiftUI + Combine, for Android — Jetpack Compose + Flow, for cross-platform — Flutter 3.x.
- Implementation — we write the MQTT client with reconnection handling using exponential backoff algorithm, data processing logic, and the dashboard. For the backend, we use Firebase or Supabase. We add deep linking to navigate to a specific device. We use
async/awaitin Swift andCoroutines + Flowin Kotlin for concurrency. - Testing — on a real device or emulator (if hardware is available). We test scenarios: connection loss, buffer overflow, malformed JSON. Per App Store Review Guidelines (Section 4.2), the app must have minimal functionality, so we always include charts and notifications.
- Deployment — publication to App Store and Google Play, with TestFlight / Firebase App Distribution setup.
Typical mistakes at the start
Common issues: no MQTT reconnection handling — the app "freezes" on disconnect; too frequent requests (less than 1 second) — the device can't respond in time; ignoring App Store requirements — without charts and notifications, the app may be rejected. We account for these nuances and design a fault-tolerant architecture from the start.
What's included in the implementation
- Dashboard with metrics (CPU, RAM, temperature, RSSI, uptime, firmware version)
- Last errors screen with filter by time and level (info/warning/error)
- Push notifications (APNs / FCM) on threshold breaches
- MQTT integration (TLS support, last will)
- API and device configuration documentation
- Source code and build instructions
- 3 months of free support after delivery (3-month warranty guarantee)
Timelines and cost
A basic version takes 1–2 weeks and starts at $3,500. If you need charts, multiple device groups, or a web admin panel — up to 4 weeks, with custom pricing around $8,000–$12,000. The cost is calculated individually after reviewing your API and requirements. Our proven experience with 30+ IoT projects guarantees reliable delivery. Our approach reduces operational costs by 30–40%, translating to $20,000–$50,000 annual savings for mid-sized deployments.
Common questions
Many users ask about metrics collection: Typically, CPU and RAM usage, processor temperature, uptime, firmware version, WiFi level (RSSI), MQTT reconnection count, and recent system log errors. The set depends on the platform – Linux, OpenWRT, ESP-IDF. Data transmission is most commonly via MQTT with a JSON payload. The device publishes structured data to a topic, and the mobile client subscribes and displays it in real time. Alternatives include REST API, WebSocket, or CoAP. Development time: a basic version with a metric dashboard and error list takes 1–2 weeks; with charts, push notifications, and backend integration, 3–4 weeks. Regarding WiFi level (RSSI) below -80 dBm, such a signal is critically weak and often causes MQTT message loss and false alarms. We recommend improving network coverage or using a wired connection for critical devices. We guarantee that the app works correctly with your equipment. If errors occur, we fix them free of charge within 3 months after project delivery.
Contact us for a consultation — we'll evaluate your project in 1 day. Reach out, and we'll show you how diagnostics can become fast and clear. Get a free analysis of your equipment and implementation recommendations.