@dsj Does WF implement any usage limits on the “New observation” trigger? I have been having a play with IFTTT today using two applets with the “New observation” trigger to write my station data to two separate Google Sheets. I have run into a Usage limit exceeded error, and was wondering if this was a WF imposed limit, or a Google Sheet imposed limit?
Hi Peter. We’re not imposing any usage limits so I suspect it’s on Google’s side.
Hi David, thanks for confirming.
On a different but related note, I have noticed some small differences between the data in the app/web compared to the data stored in the IFTTT Google Sheet. For example, between 19:36 and 19:39 today, there is a gap in the data coming from both my Air and Sky modules:
I am not entirely sure what has caused the gap (that remains to be troubleshooted), but if I examine the IFTTT Google Sheet around the same time, then the gap is shifted by a minute:
|February 5, 2019 at 07:36PM||1549395365||5.5||88||1010.9|
|February 5, 2019 at 07:37PM||1549395425||5.5||88||1010.9|
|February 5, 2019 at 07:40PM||1549395605||5.5||88||1010.7|
|February 5, 2019 at 07:41PM||1549395665||5.5||88||1010.7|
This shift appears to persist, such that right now the latest timestamp on the web/app graphs is 21:51, whereas in the Google Sheet it is 21:52.
There also appears to be a difference in the rounding between the web/app and the Google Sheet. Often the values differ by 0.1.
When I came across these limits, (Which I was surprised were imposed) I expected that if I continued with my development plans in google sheets I would be exceeding their limits on my free account.
edit: Particularly when I created the function mentioned here:
I thought that if I was calling my function (along with other future plans) I would possibly hit these limits:
So I have been playing with a Google Script to address the 2000 row limit. Each time the Google Sheet receives a new observation the scripts takes about 0.5 seconds to run. Then at midnight each day when the sheet is copied to a fixed width text file, the script runs for about 8 seconds. I don’t think I have exceeded any of the script limits yet…
You should notice this shift roughly half the time. That’s because the shift is related to the “station observation” concept and the fact that sensor observations not necessarily arrive at the same point in time. The station observation, as you know, is a “federated” observation built from one or more sensor observations at the same station. Since one observation can’t have two timestamps (well, it could, but that would be confusing), the timestamp of the station observation is whenever it’s created, using the latest data from every sensor device involved. Now, since the data pushed to IFTTT (and WU and other data integrations) is based on building a station observation on a fixed schedule (every minute), its timestamp won’t necessarily correspond to the device observation timestamp. It should always be within 30 seconds of either side, however (assuming all devices are online).
Hmm. That’s probably got something to do with the unit conversion process. We’ll have a look.
Thanks for the reports!
Yep, makes sense. Thanks for taking the time to explain.