https://oss.ecowitt.net/uploads/20260114/HTTP%20API%20interface%20Protocol%20(Generic)-(V1.0.6-2026-1-14)%20.pdf

[email protected] schrieb am Sonntag, 8. Februar 2026 um 17:08:35 UTC+1:

> Where can one find the documentation for the API(s).
> For what its worth, I am most interested in the backfill
> Thanks. rich
>
> On Sunday, 8 February 2026 at 09:19:56 UTC-5 [email protected] wrote:
>
>> OMG I just took a first glance at their "HTTP API":
>> "common_list": [{
>>             "id": "0x02",
>>             "val": "8.4",
>>             "unit": "C"
>>         }, {
>>             "id": "0x07",
>>             "val": "83%"
>>         }, {
>>             "id": "3",
>>             "val": "6.2",
>>             "unit": "C"
>>         }, {
>>             "id": "5",
>>             "val": "0.187 kPa"
>>         }, {
>>             "id": "0x03",
>>             "val": "5.7",
>>             "unit": "C",
>>             "battery": "0"
>>         }, {
>>             "id": "0x0B",
>>             "val": "6.84 km/h"
>>         }, {
>>             "id": "0x0C",
>>             "val": "10.80 km/h"
>>         }, {
>>             "id": "0x19",
>>             "val": "27.00 km/h"
>>         }, {
>>             "id": "0x15",
>>             "val": "2.29 W/m2"
>>         }, {
>>             "id": "0x17",
>>             "val": "0"
>>         }, {
>>             "id": "0x0A",
>>             "val": "340"
>>         }, {
>>             "id": "0x6D",
>>             "val": "358"
>>         }
>>     ]
>> I mean, WTF? We have HEX IDs, mixed with integer IDs, we have items with 
>> attributes for values and one for their unit here, but also have value 
>> string with units there, and then just values without unit at all, the 
>> battery indicator is tied to the computed "windchill" value (0x03) and not 
>> to a physical device. The inTemp (0x01) is not there, because it's not in 
>> the "common_list", when you have a WH25 Temp/Humi/Pressure sensor attached. 
>> Could it be worse?
>>
>> [email protected] schrieb am Freitag, 16. Januar 2026 um 21:47:07 
>> UTC+1:
>>
>>> And one could separately deal with the backfilling.
>>> Another approach would be a custom firmware for their devices, under the 
>>> hood ESP32 are doing the job. I didn't do more on this platform but 
>>> building solar powered humitemp sensors that emit MQTT readings, so that's 
>>> beyond my resources. Another more generic approach might be using something 
>>> like the LILYGP LoRa32 devices to transform the raw sensor data into MQTT 
>>> messages. Whatever would be the most straightforward way to get the sensor 
>>> readings into the loop.
>>>
>>> Vince Skahan schrieb am Freitag, 16. Januar 2026 um 21:16:38 UTC+1:
>>>
>>>> I'm sure a lot of ecowitt's oddities are them cutting corners to 
>>>> minimize cost.
>>>>
>>>> My opinion re: MQTT is that it removes a lot of complexity and results 
>>>> in a far more maintainable setup.
>>>>
>>>> If you look at the ecowitt_local_http driver it's an unsupportable mess 
>>>> to my eyes.  Thousands and thousands of lines of code to deal with the 
>>>> various sensor models.  A driver that is over 13,500 lines of code is just 
>>>> beyond too much.  Heaven knows how big it would need to grow over time as 
>>>> Ecowitt continues to add new sensors every week it seems.
>>>>
>>>> If you instead simply have the gateway emit MQTT it's much (much) 
>>>> easier to see what it's transmitting and map the items specifically in 
>>>> weewx.conf and ignore anything they emit you want to ignore.
>>>>
>>>> They even seem to emit somewhat reasonable element names in what they 
>>>> publish to MQTT, far more reasonable than what you get if you query their 
>>>> gateway directly via http://x.x.x.x/get_livedata_info which has some 
>>>> 'very' odd naming conventions if you query the gateway directly.
>>>>
>>>> On Friday, January 16, 2026 at 11:43:13 AM UTC-8 [email protected] 
>>>> wrote:
>>>>
>>>>> Will using MQTT make any real difference? Isn't it just another way to 
>>>>> get the same data in the same intervals? That's another thing Ecowitt 
>>>>> hasn't solved in the best way, imho: you can only obtain data from the 
>>>>> device at a fixed interval. Why don't they distinguish between continuous 
>>>>> data and discrete events? Why can't I receive lightning strikes as 
>>>>> distinct 
>>>>> events? Even if querying the API in a sub second interval was possible, 
>>>>> you 
>>>>> won't be able to guarantee to get a single strike event. Their console 
>>>>> with 
>>>>> the display receives and shows discrete events, their lightning strike 
>>>>> sensor emits detected strikes as they happen. On the API side you only 
>>>>> get 
>>>>> blurred-over-query-interval data.
>>>>>
>>>>> Vince Skahan schrieb am Freitag, 16. Januar 2026 um 20:29:47 UTC+1:
>>>>>
>>>>>> You're running as a driver.  You have the station type set to point 
>>>>>> to the [EcowittHttp] stanza and you selected it via 'weectl station 
>>>>>> reconfigure'.
>>>>>>
>>>>>> On Friday, January 16, 2026 at 10:32:29 AM UTC-8 NotThePainter wrote:
>>>>>>
>>>>>> On Friday, 16 January 2026 at 07:10:16 UTC-5 steepleian wrote:
>>>>>>
>>>>>> . Also the backfill only works when the extension is set up as a 
>>>>>> driver, it does not work when set up as a service.
>>>>>>
>>>>>>
>>>>>> How do I know if I running as a driver or a service, or that even not 
>>>>>> a good question.
>>>>>>
>>>>>>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/weewx-user/4bce9980-9314-4254-87a2-2f72fd77f3fcn%40googlegroups.com.

Reply via email to