Add endpoints parameter

[jsight]

No description provided.

[jsight]

This seems a little hacky to me, but it does work around the problem.

[maerco]

I tested on my deploy and it fix also my model 3 recently updated to 2023.38.4

[brontide]

This solved data loss after installation of 2023.38.4 on my Model Y.

[finnysamuel]

Thanks for the patch.. It worked - MYLR 23.38.4

[viorelovidiu]

This seems a little hacky to me, but it does work around the problem.

Hi, sorry being numb but how do I deploy this patch? I do have a Synology NAS with docker container. Thank you!

[jsight]

Change has been pushed to docker here: https://hub.docker.com/repository/docker/jsightler/teslamate/general

Also updated to (hopefully) be backward compatible and less crashy. :)

[swiffer]

@jsight query param with get instead of encoding the params yourself

See example usage here

https://github.com/elixir-tesla/tesla?tab=readme-ov-file#direct-usage

[blexter]

@adriankumpf
I would really appreciate if this PR could be approved and merged into the official release. 🙏🙏🙏

[jsight]

@swiffer Good idea, thank you!

[marcbaier]

would be great if someone could announce it here when the update is available in the Teslamate:latest version. thanks so much !

[marcbaier]

Change has been pushed to docker here: https://hub.docker.com/repository/docker/jsightler/teslamate/general

Also updated to (hopefully) be backward compatible and less crashy. :)

If I would like to use this update / patch and install it on my docker installation on my synology nas, can I just install it in docker by selecting my current teslamate/latest container and then selecting "import" "add from URL" ?

[blexter]

@jsight
Getting "Internal Server Error" regularly since updated to this fix.

[brontide]

@blexter You don't indicate if you are seeing that error in the browser or the logs. This code change is tiny and should have zero impact on the front-end of Teslamate where you might see errors in the browser. If it's in the logs please post the complete error messages so that they can be reviewed.

[blexter]

@blexter You don't indicate if you are seeing that error in the browser or the logs. This code change is tiny and should have zero impact on the front-end of Teslamate where you might see errors in the browser. If it's in the logs please post the complete error messages so that they can be reviewed.

Checked the logs, not 100% sure what's sensitive data, so I might removed "to much", tell me if so and I can somehow send missing info.

2023-11-01 09:02:12.741 [error] Task #PID<0.3967.0> started from TeslaMate.Vehicles terminating
** (KeyError) key :drive_state not found in: %TeslaApi.Error{
reason: :unknown,
message: "{"error": "Get \"http://localhost:7654/vehicle_data?slow_query_time=15&headers%5BX-Request-Timeout%5D=5&let_sleep=true&user_command=true\\\": EOF"}",
env: %Tesla.Env{
method: :get,
url: "https://owner-api.teslamotors.com/api/1/vehicles/1234567890/vehicle_data",
query: [],
headers: [
{"x-xss-protection", "1; mode=block"},
{"x-txid", "1234567890"},
{"content-type", "application/json; charset=utf-8"},
{"x-request-id", "1234567890"},
{"content-length", "219"},
{"x-envoy-upstream-service-time", "1336"},
{"date", "Wed, 01 Nov 2023 08:02:12 GMT"},
{"server", "envoy"},
{"x-frame-options", "DENY"},
{"x-content-type-options", "nosniff"},
{"strict-transport-security", "max-age=31536000; includeSubDomains"},
{"cache-control", "no-cache, no-store, private, s-max-age=0"}
],
body: %{
"error" => "{"error": "Get \"http://localhost:7654/vehicle_data?slow_query_time=15&headers%5BX-Request-Timeout%5D=5&let_sleep=true&user_command=true\\\": EOF"}",
"error_description" => "",
"response" => nil
},
status: 540,
opts: [
access_token: "abcdef1234567890"
],
module: TeslaApi,
client: %Tesla.Client{fun: nil, pre: [], post: [], adapter: nil}
}
}
(teslamate 1.28.0-dev) lib/tesla_api/vehicle.ex:59: TeslaApi.Vehicle.get_with_state/2
(teslamate 1.28.0-dev) lib/teslamate/api.ex:299: TeslaMate.Api.preload_vehicle/2
(elixir 1.15.6) lib/task/supervised.ex:101: Task.Supervised.invoke_mfa/2
(elixir 1.15.6) lib/task/supervised.ex:36: Task.Supervised.reply/4
Function: &:erlang.apply/2
Args: [#Function<2.69151209/1 in TeslaMate.Api.handle_result/3>, [%TeslaApi.Vehicle{id: 1234567890, vehicle_id: 1234567890, vin: "1234567890", tokens: ["1234567890", "1234567890"], state: "online", option_codes: [""], in_service: false, display_name: "Model S", color: nil, calendar_enabled: true, backseat_token: nil, backseat_token_updated_at: nil, api_version: 67, charge_state: nil, climate_state: nil, drive_state: nil, gui_settings: nil, vehicle_config: nil, vehicle_state: nil}]]
2023-11-01 09:02:18.157 [notice] Application teslamate exited: shutdown
Kernel pid terminated (application_controller) ("{application_terminated,teslamate,shutdown}")

Crash dump is being written to: erl_crash.dump...done

[DrMichael]

@adriankumpf Given, that without that PR teslamate is broken, we need this PR incorporated. If you don't have the time, your opinion in #3384 would be very much appreciated.
We don't probably jump to a forked teslamate.

[tpd-opitz]

How can I apply the change locally to my teslamate instance until the this PR is merged?
I thought of adding RUN directive to my dockerfile to apply the change using sed like suggested [here] (https://stackoverflow.com/a/27714071) but I cant find the file vehicle.ex within the container...

[sargue]

How can I apply the change locally to my teslamate instance until the this PR is merged? I thought of adding RUN directive to my dockerfile to apply the change using sed like suggested [here] (https://stackoverflow.com/a/27714071) but I cant find the file vehicle.ex within the container...

Well, you have several options.

The simple one is to trust the patched docker image that the same author of this PR uploaded to Docker Hub.

Another option is to build yourself a patched image, which seems to be what you're trying.

NOTE: I haven't tried any of this myself, my car is still on a previous version. Be very careful, do a full backup before you change anything.

If you have access to the command line and docker, and you have git installed (or you can install it), you can just build a local image with the change, then change the docker-compose.yml to point, temporarily, to this new local image.

It should be something like this. You can do this on the same system you're running Teslamate. Make sure you don't have already a teslamate directory because the first command will create one.

git clone https://github.com/jsight/teslamate.git
cd teslamate
git reset --hard 8134150b247eec5a08572e62fdc4ebe6fda22f63
git clean -df
docker build . -t teslamate-patch

If all commands are successful, then you can just point your docker-compose.yml to your local patched image.

services:
  teslamate:
    image: teslamate-patch

Another option is to clone the official repository, then apply manually the change, then build the image in the same way. It should be the same.

[SebastianGode]

I guess once @adriankumpf gets his car updated to the 2023.38.x this PR will get merged.
So shouldn't take that much longer again although I personally would suggest Adrian to add a second trustworthy maintainer to the repo :)

[tpd-opitz]

If all commands are successful, then you can just point your docker-compose.yml to your local patched image.

 services:
     teslamate:
         image: teslamate-patch

Is teslamate-patch just an identifier I need to use at both places, the docker build command and the docker-compose.yml or are there any constraints to the directories involved?

[sargue]

Correct, it's just a tag, any string will do. As you point out, you just need to make sure it's the same on the docker build command and the docker-compose.yml.

There's no restriction on where you build or use the image, it's all a reference in the docker local repository. You can see that with docker image ls.

By the way, I've been told that grafana needs to be updated as well, as the current master branch of teslamate has some changes as well.

If someone tries it, please share the outcome here.

[tpd-opitz]

I got it runnig:

$ docker-compose up -d
Creating network "teslamate_default" with the default driver
Creating teslamate_mosquitto_1 ... done
Creating teslamate_grafana_1   ... done
Creating teslamate_teslamate_1 ... done
Creating teslamate_database_1  ... done

$ docker ps
CONTAINER ID   IMAGE                      COMMAND                  CREATED          STATUS                   PORTS                                       NAMES
4d6aa4bdecf0   postgres:13                "docker-entrypoint.s…"   10 minutes ago   Up 10 minutes            5432/tcp                                    teslamate_database_1
4ae87db8d2a0   teslamate/grafana:latest   "/run.sh"                10 minutes ago   Up 10 minutes            0.0.0.0:3000->3000/tcp, :::3000->3000/tcp   teslamate_grafana_1

7c5632baf27d   teslamate-patch:latest     "tini -- /bin/sh /en…"   10 minutes ago   Up 10 minutes            0.0.0.0:4000->4000/tcp, :::4000->4000/tcp   teslamate_teslamate_1

8ff0a3268b98   eclipse-mosquitto:1.6      "/docker-entrypoint.…"   10 minutes ago   Up 10 minutes            0.0.0.0:1883->1883/tcp, :::1883->1883/tcp   teslamate_mosquitto_1
90a26f4c4145   buanet/iobroker            "/bin/bash -c /opt/s…"   22 months ago    Up 11 months (healthy)   0.0.0.0:8081->8081/tcp, :::8081->8081/tcp   iobroker

$ docker-compose logs -f  teslamate 
Attaching to teslamate_teslamate_1
...
teslamate_1  | 2023-11-09 22:27:18.957 [info] GET /
teslamate_1  | 2023-11-09 22:27:19.051 [info] Sent 200 in 93ms
teslamate_1  | 2023-11-09 22:35:49.847 car_id=1 [info] Start / :online
teslamate_1  | 2023-11-09 22:35:49.873 car_id=1 [info] Logged missing software update: 2023.38.6 c1f85ddb415f
teslamate_1  | 2023-11-09 22:35:49.988 car_id=1 [info] Connecting ...

And at least it updated the SoC and the firmware version (got update 2023.38.6 last night)

[sargue]

@tpd-opitz , remove the :latest from the image line on the docker-compose.yml. It's trying to pull the image from a remote repository, but there's no remote repository (the image is local).

[brianmay]

Big question is: Is this change permanent? Or is Tesla going to change in the future and break this yet again?

[jaymunro]

Is there some way to up the priority on this? TeslaMate is completely broken with the the new API from Tesla.

[nelsongraca]

any timeline to merge this?

[jsight]

@brianmay It is in the official API documentation. They could change it further, but I'd guess they will document any further changes as well.

[brianmay]

There is official API documentation? Where is this? Everything I have seen clearly says unofficial only.

[jsight]

@brianmay Official API docs: https://developer.tesla.com/docs/fleet-api#vehicle_data

It is listed as the "fleet API", but the ones that I've seen are the same ones that we are used to. The notes call out the specific changes in 2023.38+ as well.

[swiffer]

Fact is that the current latest version 1.27.3 is broken and the fix discussed here is working well (and is documented change for the fleet API). Not having this or #3398 merged means broken Teslamate. There are more and more forks popping up

• https://github.com/MyTeslaMate/teslamate
• https://github.com/teslamate-fork/teslamate-fork

In addition there are efforts of coordinating a fork

• TeslaMate - Still active? or Dead? #3396
• Broaden the administration? #3384
• Is this dead? Pressing issues and PRs remain unapproved or unreviewed for months. #3399

It would be really great to avoid that, broden the admin and combine forces in making teslamate a better product - or at least try to coordinate forking.

[brianmay]

I have had to use my fork for some time, simply because I have not had any feedback on my pr #3087. This in turn reduces motivation to work on other issues that have been nagging me.

I can understand though that @adriankumpf may feel that he is unable to keep with with this project anymore. I have been in this situation myself on a number of occasions.

I think we are all very grateful for the time and effort that @adriankumpf and others have spent on this project to date.

Handing over the project is a big step too, can you really trust the new maintainer? I have heard some horror stories where this can go wrong. In at least one case the new maintainer inserted malicious code into a project he had just taken over. And plenty of times where I accidentally end up being the new maintainer, and I am already over committed.

The problem with creating a fork, is that 50 forks is not going to help. We need one fork that the entire community can get behind and support. Maybe one, like the above that isn't directly tied to one person, so that the responsible people can change with time too.

If we can agree on one fork to be the new official repo, we can then ask github to mark it as such. Then the annotations will change to make this project an fork of the new official fork, rather then vice verse.

[szinn]

I can confirm this API works on 2023.27.7 as well. The vehicle_data_combo likely isn't needed since it appears to be an encoded blob and doesn't add anything new to the remainder of the request.

[Dulanic]

With #3398 merged, this is is no longer needed correct?

[swiffer]

@Dulanic - yes, can be closed

[Dulanic]

I have had to use my fork for some time, simply because I have not had any feedback on my pr #3087. This in turn reduces motivation to work on other issues that have been nagging me.

I can understand though that @adriankumpf may feel that he is unable to keep with with this project anymore. I have been in this situation myself on a number of occasions.

I think we are all very grateful for the time and effort that @adriankumpf and others have spent on this project to date.

Handing over the project is a big step too, can you really trust the new maintainer? I have heard some horror stories where this can go wrong. In at least one case the new maintainer inserted malicious code into a project he had just taken over. And plenty of times where I accidentally end up being the new maintainer, and I am already over committed.

The problem with creating a fork, is that 50 forks is not going to help. We need one fork that the entire community can get behind and support. Maybe one, like the above that isn't directly tied to one person, so that the responsible people can change with time too.

If we can agree on one fork to be the new official repo, we can then ask github to mark it as such. Then the annotations will change to make this project an fork of the new official fork, rather then vice verse.

I dont disagree, there needs to be a head person responsible for all of it. Who that can be? I'm not sure at this point. But I think we are open to talking about it. There is a open discussion currently.