Iot hub message quota increasing out of nowhere

Hi @BPS Sistemas de Automação Greetings! Thank you for posting the question here. Could you please confirm if there are any Edge devices in your IoT Hub?

I have encountered the following related Q&A issues that are reported in the past which state the same issue.

• Total Number of messages used in IoT Hub is increasing even if no Telemetry message is sent
• Azure IOT Hub: messages send when device is doing nothing

Both issues indicate that there are Twin reads performed by the IoT Edge modules are contributed to the total number of messages. These messages (mostly GetTwin and update Reported properties in your case) are part of normal IoT Edge device operations and they can be tweaked at the cost of slower responsiveness from the Edge device by using the Env Variables (especially MinTwinSyncPeriodSecs and ReportedPropertiesSyncFrequencySecs

Can you please confirm if that is applicable to you?

PS 3: From time to time I have problems with the connection failure of the HUB with some devices.

As far as I know, IoT hub has a limitation of Total number of messages/day per IoT Hub unit which is 8000 for free tier. It should be able to handle 70 devices without any issues.

Do you notice the devices reconnect after the connect failure? I appreciate it if you can share more details on the protocol/SDK the devices use and the frequency at which they transmit the data to better understand and assist you with the issue.

Hi @LeelaRajeshSayana-MSFT, thanks for the feedback.

Answering the first question: There are no EDGE devices currently on my IoT HBU, just regular devices.

I read the questions you mentioned but they don't apply to me. Regarding DevicesTwin, I perform some DeviceTwin readings in a web app, but they are at most 3 times a day, it would not justify such a high increase.

Regarding the message quota, when the device fails to connect to the HUB, the message quota has not yet been reached.

Device refresh rates are as follows:

• Twice a day, more specifically at 07:30AM and 13:00PM, the telemetry is sent to the hub. A json containing 20 keys with their respective values.
• Every hour during business hours (07:30AM to 18:00PM) DeviceTwin reported messages are updated. A json containing 6 keys with their respective values.

Regarding the architecture of IoT devices:

My devices are based on ESP32 WROOM, they are programmed using Arduino Framework. I'm currently using a library for esp32 and IOT HUB that I found on GITHUB (Available HERE). This library is discontinued. But I'm having a lot of problems with ESP32 and this library, the device restarts a few times throughout the week and sometimes the device just freezes, and I believe that the problem could be this library. In this way, I converted my code to use the official Azure SDK for Embedded C for Arduino library (Available HERE), provided and maintained by Microsoft. However, I haven't updated the firmware of the devices yet, I have only 1 device with this library under test in the same HUB. As soon as the trial period passes I will update all devices with this new library.

Hello @BPS Sistemas de Automação,

you mention you use the IoT Hub free tier.

There are some limitations compared to the other tiers except for the quality of the service.

Next to the limited number of messages it can receive, the size of each message is also smaller.

The Free tier IoT Hub can receive 8000 'chunks' per day.

One message from your device is divided into these chunks.

Each chunk has a size of 0.5KB.

So, if you send a message of 2.25KB, that is counted as 5 chunks!

Please check the message size your devices are sending.

This could also explain the number of IoT Hub messages consumed.

Notice, the S1 tier IoT Hub starts with 400.000 chunks per day with a size of 4KB.

Good morning,

Today I updated the firmware of the other 70 devices with the new azure iot library and to my surprise the devices have many connection failures. When executing the direct methods some do not respond. Could this be something related to the hub? or the network infrastructure? The problem happens in more than one place, that is, even with different infrastructures the problem is the same.