buck doctor

The purpose of this command is to help the user debug and solve and issue of another buck command. This command requires implementing custom server-side support to be fully functional. We do not provide a functional implementation with buck, however we have documented the protocol and theDoctorCommandIntegrationTest contains very simple examples of interactions between the doctor command and server.

You can use this command after setting up the environment.

buck doctor

Command flow

The command supports the debugging of other commands except invocations of buck rageand buck doctor. At the start the user has to choose the buck invocation that they want to debug. In order for a buck invocation to appear on the list it must have log files in buck-out. After choosing the command a report is generated and either stored locally or is uploaded to a remote endpoint defined under [rage] in buckconfig. Also, another Json report is generated structured after class AbstractDoctorEndpointRequest and is send to the remote endpoint defined under [doctor] in buckconfig. This report also contains a URL to the location of the rage report. Then, the doctor remote endpoint produces a response that contains information about the status of core components like parsing, environment and remote cache along with suggestions on how to solve potential problems.

Setup notes

• The command requires the creation of at least one remote endpoint that handles the Json request and creates the fixing suggestions based on the Json format ofAbstractDoctorEndpointResponse.
• If you want to also have the data from the creation of the rage report you will need to define another remote endpoint that can get the information, otherwise the zipped report will be saved on disk
• All the communication is sent using POST.

Formats & Structures

• Request is based on the Json format of class AbstractDoctorEndpointRequest.
• Response is based on the Json format of class AbstractDoctorEndpointResponse.
• The rage report zip file contains a Json file will metadata and useful information along with a list of log files and local configuration.

You can pass extra arguments to the doctor request by defining the [doctor_extra_request_args]. This will add them as extra parameters in the post request. In this case the main information will be under parameter data.

Buckconfig example

[rage]
    report_upload_url = https://localhost:4546
    report_max_size = 100MB

[doctor]
    endpoint_url = https://localhost:4545
    extra_request_args = ref=>12ab,token=>42

Request example

{
    buildId: "...",
    rageUrl: "...",
    machineReadableLog: "...",
}

Response example

{
    errorMessage: "...",
    suggestions: [
                   "state": "ERROR",
                   "area": "Environment",
                   "suggestion": "Parsing too slow. Check BUCK file.",
                 ]
}