Too many requests in error log

[nagtegaal]

Deze fout is ontstaan door een aangepaste integratie.

Logger: custom_components.parcelapp.coordinator
Bron: helpers/update_coordinator.py:435
integratie: Parcel App (documentatie, problemen)
Eerst voorgekomen: 9 november 2025 om 16:22:08 (38 gebeurtenissen)
Laatst gelogd: 05:22:09

Error fetching parcelapp data: Error fetching data from API: 429, message='Too Many Requests', url='https://api.parcel.app/external/deliveries/?filter_mode=recent'

---

I get several of these errors a day, and would like to fix these. How should this be done?
I think that the current interval for polling is set to 5 min.
What service should I use to set the polling manually to - as example 10 min - and fix the above errors in my logs? Should that work?

[jmdevita]

@nagtegaal right now, the 5 minute interval is hard-coded in the integration itself. But the max amount of API requests we can ping the server is 20/hr. I'm surprised you're hitting the limit though, as there should be a 40% buffer (the integration only pings 12/hr).

An option is to have a configurable field of API requests/hr in the integration, but I'm curious how you ended up hitting the limit. Do you have multiple integrations with the same API key?

[nagtegaal]

I do not have multiple integrations with the same api key running, but I do have multiple cards using the integration in the same dashboard page. (In a bubble pop-up card, bubble cards and inside a raw yaml-card)

Wijle I do not thing that coupe be the cause of this error, perhaps it is related. Cant think of something else…

[jmdevita]

@nagtegaal that is quite odd.. Multiple cards using the integration shouldn't cause any issues. Are you using the API key outside of home assistant with anything else? And do you constantly run into this issue with Error 429 or is this every once in a while?

Actually, I see it in my home server as well and I know I'm not using it anywhere else. I'll look into it, that is really weird.

[jmdevita]

@nagtegaal found it! turns out each sensor was calling the coordinator, not the other way around. So there are times where the sensors are not synced correctly, so then each sensor calls the API independently (so instead of 12x/hour, it could try up to 48x/hour). I instead changed this behavior and to only receive from the coordinator (which calls 12x/hour).

I have tested locally, but if you want to try with the branch in this thread, go for it! I have released a beta version with this fix here. You can go to HACS, click "redownload" and select the beta version if you're interested.

cc @PineappleEmperor lmk your thoughts ^

[nagtegaal]

After updating and restart of HA (which it suggested after the update of the parcel integration):

So... It isn't fixed. It f#cked up completely and doesn't work at all anymore...

[nagtegaal]

Hmmm... After 20 minutes the entities seem to be available again, while I was expecting that (according the update interval of 5 minutes) this was going to be a lot sooner...
So far, I do not have the "Too many requests" in my error-log file.

[jmdevita]

@nagtegaal I'm guessing what happened is that you exhausted your hourly limit, so when you restarted and applied the patch on the beta branch, it couldn't validate because you used up your hourly quota. After 20 minutes your quota reset and the integration worked.

Can you confirm if you're still not receiving any issues on your log file? So far, I'm not seeing any on my instance.

[nagtegaal]

I had one more error 429 error overnight after installing the beta, which is already a lot better than it was.
It's at a time I was asleep, so the cause of the error is really by the integration itself/home assistant.

If the parcel-server is unavailable, is the integration hammering until it gets a response? That could also be a reason why it is making too many requests to the parcel-server. Although, I'm not fully into the inner workings of HA itself to know how this is handled, nor the inner workings of the integration once the parcel-api-server is down/unreachable.

I'll wait a bit more, and report back to you tomorrow.

[PineappleEmperor]

@jmdevita sorry for the delay! Just had a look through and your updates match more closely what I've done/seen in other integrations now.

Just checking the last image from @nagtegaal though, I think the integration will only pull the recent list, which won't fully populate the other sensors - might need to rework the recent sensor to take a portion of the active or full parcels to compensate?

[jmdevita]

@PineappleEmperor No prob at all! Thanks for reviewing :) What are you thinking with reworking the recent sensor? I wasn't quite following.

Also, @nagtegaal that's a good question. Right now with this change, the coordinator just pings the server every 5 minutes, even if the server is down, errors out, succeeds, etc. So, it won't rapidly ping the server when it's down. I'll continue investigating. May be worth having the interval configurable on how much it scans per hour.

[ozmozi]

Just another idea: if you make the update interval configurable (which would be a nice feature), could you also add something (binary sensor?) that could end up being linked to a button on a dashboard to press to initiate an update right now? It could then also be used in automations. Sometimes I just hate to wait for 5 minutes ha ha.

Some info on how many API calls have been made in the last hour (?) would be useful too in this context but that is not at all essential.

[PineappleEmperor]

@jmdevita it's around whether you request 'recent' parcels, which if memory serves provides a somewhat random selection of parcels, vs 'active', which provides everything:

Endpoint
URL: https://api.parcel.app/external/deliveries/

Path Parameters
filter_mode (optional).
Possible values: "active", "recent". Default": "recent".

Example Request
curl "https://api.parcel.app/external/deliveries/?filter_mode=active" -H "api-key: YOUR_API_KEY" 

I might be misremembering what the endpoint returns though!

@ozmozi it may be possible to add a service that provides a manual refresh? Currently you could achieve this by reloading the integration, since that causes it to do an initial API call.

As for the API calls in the last hour, this would be from the perspective of the integration, which would get reset if you restarted, or reloaded the integration 😅 It's still possible to attempt, but we disregarded it early on I think.

[ozmozi]

@PineappleEmperor good idea to re-download the integration. I’ll try that next time I get impatient.

We did discount the idea of counting the API calls/hr earlier, but I think that was mainly because it was fixed at every 5 minutes, or 12/hr. I think there could be some value in having the counter if the timeframe becomes variable or adjustable. However it is far from essential. And one would have to live with the counter resetting if HA resets etc. That’s one of my pet hates about HA in general, as I can’t be bothered to implement the work-arounds that do exist to have more “permanent” variables.

[PineappleEmperor]

Just to be clear - I meant reload the integration! You should be able to grab the yaml from the developer actions menu:

action: homeassistant.reload_config_entry
data:
  entry_id: <select in the UI the parcelsha integration>

Yeah exactly, the counter would be unreliable in a lot of situations where you'd potentially be getting near the limit due to additional queries from reloads, restarts, etc.

[ozmozi]

Thanks for the clarification 😁