Azure IOT Hub - Message Per second throttle limit

Question

Azure IOT Hub - Message Per second throttle limit

Alen Mathew 0

I need get a clarity on message that can be throttled per second to IOT hub without causing throttling error 429

Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-08T18:36:13.9966667+00:00
Hello @Alen Mathew

It looks like you're trying to understand the device-to-cloud message rate at which Azure IoT Hub starts throttling (HTTP 429). The limit depends on your tier and the number of units allocated.

Here’s a simplified and accurate breakdown:

• Free and S1 tiers

Throttling is based on total operations per unit, not a simple per-second formula

S1 provides up to 12 million messages per day per unit, which roughly translates to about ~140 messages/sec per unit (average rate)

There are also per-minute and per-second throttling protections, so short bursts beyond this rate may result in 429 errors

• S2 tier

Approximately 1200 messages/sec per unit

• S3 tier

Approximately 6000 messages/sec per unit

Important notes:

IoT Hub uses traffic shaping, meaning short bursts may be allowed, but sustained traffic above the limit will trigger throttling (HTTP 429)

Limits apply across all device-to-cloud operations (telemetry, twin updates, etc.), not just telemetry messages

Recommendations to avoid throttling:

Implement retries with exponential backoff and jitter in your device/application logic

Monitor metrics such as Throttled Requests in Azure Monitor

Scale your IoT Hub by adding more units if you need higher sustained throughput

References:

IoT Hub quotas and throttling

IoT Hub traffic shaping

Monitor IoT Hub metrics

I Hope this helps. Do let me know if you have any further queries.

Thankyou!
Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-09T20:10:15.77+00:00

Hello @Alen Mathew

Did you get any chance to review the response.

Thankyou!
Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-10T15:11:04.8266667+00:00

Hello @Alen Mathew

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Thankyou!

1 answer

Your answer

Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-08T18:36:13.9966667+00:00

Hello @Alen Mathew

It looks like you're trying to understand the device-to-cloud message rate at which Azure IoT Hub starts throttling (HTTP 429). The limit depends on your tier and the number of units allocated.

Here’s a simplified and accurate breakdown:

• Free and S1 tiers

Throttling is based on total operations per unit, not a simple per-second formula

S1 provides up to 12 million messages per day per unit, which roughly translates to about ~140 messages/sec per unit (average rate)

There are also per-minute and per-second throttling protections, so short bursts beyond this rate may result in 429 errors

• S2 tier

Approximately 1200 messages/sec per unit

• S3 tier

Approximately 6000 messages/sec per unit

Important notes:

IoT Hub uses traffic shaping, meaning short bursts may be allowed, but sustained traffic above the limit will trigger throttling (HTTP 429)

Limits apply across all device-to-cloud operations (telemetry, twin updates, etc.), not just telemetry messages

Recommendations to avoid throttling:

Implement retries with exponential backoff and jitter in your device/application logic

Monitor metrics such as Throttled Requests in Azure Monitor

Scale your IoT Hub by adding more units if you need higher sustained throughput

References:

IoT Hub quotas and throttling

IoT Hub traffic shaping

Monitor IoT Hub metrics

I Hope this helps. Do let me know if you have any further queries.

Thankyou!
Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-09T20:10:15.77+00:00

Hello @Alen Mathew

Did you get any chance to review the response.

Thankyou!
Anshika Varshney 13,040 Reputation points Microsoft External Staff Moderator

2026-06-10T15:11:04.8266667+00:00

Hello @Alen Mathew

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Thankyou!

Answer 1

IoT Hub throttling is governed by operation-specific limits that are applied per hub (and in some cases per hub per unit) and per second. To avoid 429 throttling errors, the total rate of each operation type must stay within its corresponding throttle for the SKU and number of units.

Key points from the throttling model:

Operation throttles are rate limits
- Throttles are enforced as rate limitations (per second, often also evaluated over minute ranges) and are subject to traffic shaping.
- When limits are exceeded for too long, IoT Hub responds with HTTP 429 and an IoT Hub error code such as 429001 ThrottlingException.
- Clients should implement back-off and retry when receiving 429 responses. See Retry patterns.
Device-to-cloud (D2C) and cloud-to-device (C2D) message rates
- D2C and C2D throttles define the maximum rate at which messages can be sent, independent of the 4‑KB metering chunks.
- D2C messages can be up to 256 KB; C2D messages can be up to 64 KB. These are maximum message sizes, not per-second limits.
- To avoid throttling, the total D2C send rate from all devices and the total C2D send rate from the service must remain below the per-hub limits for the chosen tier and unit count.
Example of traffic shaping and 429 behavior
- If an S1 hub with a limit of 100 D2C messages/sec receives 200 messages/sec:
  - Initially, IoT Hub accepts and processes the burst, queuing excess requests.
  - Over time, IoT Hub processes only 100 messages/sec and latency increases.
  - When the backlog queue fills, IoT Hub starts returning 429 ThrottlingException.
- This behavior applies similarly to other throttled operations.
Direct methods and jobs – bandwidth-based throttling
- For direct methods, throttling is based on bandwidth per unit:
  - Payload 0–4 KB counts as 4 KB; up to 40 calls/sec/unit before hitting 160 KB/sec/unit.
  - Payload 4–8 KB counts as 8 KB; up to 20 calls/sec/unit.
  - Payload 156–160 KB counts as 160 KB; only 1 call/sec/unit.
- For S3 tier jobs (update twin, invoke direct method via jobs), there is a 50 jobs device operations/sec/unit limit; direct method calls outside jobs still use the 24 MB/sec/unit limit.
Device connections throttle (separate from message rate)
- Device connections have their own throttle (for example, 100 new connections/sec for a single S1 unit), independent of the maximum number of simultaneously connected devices.
- This affects how quickly devices can connect, not how many messages per second they can send once connected.
Monitoring and tuning to avoid 429
- Use the Telemetry message send attempts metric and compare it with the documented throttling limits to see if the hub is being driven near or beyond its capacity.
- Monitor the Number of throttling errors metric, which tracks 429001 ThrottlingException.
- If throttling persists even after reducing message rates and implementing back-off, consider scaling up the IoT hub (more units or higher tier).
Error codes related to throttling
- 429000 GenericTooManyRequests: too many requests overall; reduce request rate.
- 429001 ThrottlingException: operation-specific throttling; see IoT Hub quotas and throttling.
- 429002 ThrottleBacklogLimitExceeded: backlog queue exceeded; reduce rate and review traffic shaping behavior.
- 429003 ThrottlingBacklogTimeout: backlogged requests timed out; reduce rate.
- 429005 DeviceThrottlingLimitExceeded: per-device throttling; reduce that device’s request rate.

To determine the exact “messages per second” value that avoids 429 for a specific hub, consult the IoT Hub quotas table for the chosen tier and number of units, then ensure total D2C and C2D message rates stay below those per-second limits, factoring in traffic shaping and backlog behavior.

References:

Share via

Azure IOT Hub - Message Per second throttle limit

1 answer

Your answer