Tesla Fleet API usage jumped from 40 to 5,000 requests a day with no config change

Posting this partly to ask if anyone else has seen it, partly as a data point.

I have one Powerwall site and two cars on the Tesla Fleet integration, using my own developer app. Until around 25 July my Tesla developer dashboard showed roughly 40 billed requests a day. From 3 August it went to 2,400, then 2,600, then 5,100 in a day, and I got the 80% billing limit email. At that rate it’s about £160 a month.

Nothing changed at my end. I have InfluxDB history going back, and the integration was polling live_status every 60 seconds on 10 July as well, on days when Tesla counted about 40 requests. Same entities, same credentials, same intervals either side of the jump. The client behaviour looks identical and only the counting changed.

The integration docs say “Energy product APIs are free to use”, and I assume that’s why the energy coordinators run at 60 seconds while vehicles are at 10 minutes. live_status and site_info together are 2,880 requests a day, so if energy has started being billed it would hit every Powerwall owner using this integration.

Has anyone with a Powerwall seen their Fleet API usage or bill jump in the last couple of weeks? Trying to work out whether this is specific to my account or the start of something wider.

I’ve applied the custom polling interval workaround, polling off in system options and update_entity on a 15 minute schedule, and that’s brought it down fine. I’d still like to understand what changed.

HA 2026.7.4.

Problem sorted, nothing to do with HA, it was Teslamate polling on the API.