Repository navigation
Add AWS IoT MQTT to resolve 429 Rate Limit issues and provde real-time device shadow updates - #21
Merged
Merged
Conversation
Contributor
Author
|
@benjycov ack this is a pretty large PR; let me know if you'd prefer me to split it up into smaller chunks. |
abrainwood
changed the title
Replace REST polling with MQTT push as the primary data channel for eXO devices. The Zodiac login response includes AWS temporary credentials (previously discarded) which allow connecting to the AWS IoT shadow endpoint via MQTT over WebSocket with SigV4 auth. New module mqtt_client.py: - ExoMqttClient wrapping awsiotsdk with CRT-to-asyncio thread bridge - Connect, subscribe to shadow topics, publish desired state - Automatic reconnection with re-subscribe on resume - 15-minute heartbeat (shadow/get) to keep data fresh during quiet periods - Reconnect failure detection triggers credential refresh Changes to api.py: - _full_login and _refresh_token extract AWS credentials from response - Coordinator fed by MQTT push via async_set_updated_data() - MQTT-first startup: initial data loaded via MQTT shadow/get, avoiding REST data fetch and the 429 rate limits it triggers - REST polling relaxed to 1-hour fallback (resets on each MQTT push) - Writes route through MQTT publish when connected, REST POST as fallback - Credential refresh scheduled as background task before expiry - Reconnect-failure handler refreshes credentials and reconnects - Proper cleanup in async_unload_entry (MQTT disconnect, task cancellation) Other changes: - manifest.json: add awsiotsdk dependency, iot_class changed to cloud_push - number.py: fix _attr_step -> _attr_native_step (fixes pH validation bug) - __init__.py: call cleanup_entry on unload Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- scripts/dev-setup.py: auto-onboards HA, configures debug logging, adds exo_pool integration via config flow using .env credentials - docker-compose.dev.yml: isolated dev HA instance on port 8125 - Makefile: dev, stop, logs, restart, test targets - README: add What's New entry for MQTT, update Features/Limitations, add Development section with quick-start guide - .gitignore: exclude .env and .dev-token Usage: create .env with EXO_EMAIL/EXO_PASSWORD, then make dev Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
37 tests covering: - ExoMqttClient: connect, disconnect, subscribe, publish, reconnect, heartbeat lifecycle, reconnect failure callback, credential refresh - Shadow data contract: real device data verified against all entity read paths (sensor, binary_sensor, switch, number, climate) - Write payload shapes: pool, heating, schedule writes - Inverse validation: update/accepted and update/delta do NOT feed coordinator (only update/documents and get/accepted do) - Thread bridge: CRT callbacks are bridged to HA event loop Tests are isolated from Home Assistant via conftest.py module stubbing. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Owner
|
Wow amazing work!!! Thank you so much... I'm going to test against my own Exo before merging in, will do this now! |
Contributor
Author
|
No probs! Thanks for all the work you've put into the integration - it's excellent! |
Owner
|
Hi Andrew, this is all live now in v0.2.0 - thank you so much for your work on this. I can't believe it never occurred to me to use the MQTT feed! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR replaces REST polling with MQTT push as the primary data channel, using the same AWS IoT Device Shadow protocol that the official iAqualink iOS app uses. The result is sub-second state sync, zero 429 rate limit errors, and instant writes.
No breaking changes. No additional setup required - the integration uses AWS credentials already returned by the Zodiac login API (previously discarded). Existing config entries work unchanged. REST polling is kept as a 1-hour fallback.
What changed
mqtt_client.py: wrapsawsiotsdkto connect to AWS IoT via MQTT over WebSocket with SigV4 auth. Handles subscribe, publish, reconnect, heartbeat (15-min shadow/get), and reconnect failure recovery.api.py: login/refresh now extracts AWS credentials from the response. Coordinator fed by MQTT push viaasync_set_updated_data(). Writes route through MQTT when connected, REST as fallback. Credential refresh scheduled before expiry. Proper cleanup on unload.manifest.json: addedawsiotsdkdependency,iot_classchanged tocloud_push.number.py: fixed_attr_step->_attr_native_step(fixes pH set point validation error in HA frontend).__init__.py: callscleanup_entryon unload (MQTT disconnect, task cancellation).docker-compose.dev.yml,Makefile,scripts/dev-setup.pyfor contributor onboarding.How it works
update/documents,get/accepted, etc.)coordinator.async_set_updated_data()-> entities update instantlyshadow/updatevia MQTT (REST fallback if disconnected)Testing
Test plan
🤖 Generated with Claude Code with guardrail rules for strict TDD and clean code.