Add OTA failure reason reporting and extensible status details

[MathiasKoch]

Summary

Adds granular failure diagnostics to OTA job status updates. When an OTA operation fails, the job status now includes a human-readable reason string and numeric error_code in the status details sent to AWS IoT — enabling cloud-side root cause analysis without device log access.

Introduces the StatusDetailsExt trait, allowing PAL implementations to inject custom key-value pairs (device temperature, battery level, firmware version, etc.) into every job status update alongside the base OTA fields.

Also adds self-managed integration test infrastructure that automates OTA job provisioning, S3 upload, and cloud-side assertion against AWS IoT.

Note: This is PR 3 of 4 in a stacked series. Depends on #83 (generic MqttClient trait).

Design

Failure Reason Flow

OtaError::Pal(pal_err)
  → ImageStateReason::Pal(OtaPalError)
    → JobStatusReason::Aborted(Some(reason))
      → OtaStatusDetails { reason: "sig_check_failed", error_code: 1001 }
        → MQTT job update with status details

OtaPalError variants map to reason strings and error codes:

• SignatureCheckFailed → "sig_check_failed" / 1001
• FileWriteFailed → "file_write_failed" / 1002
• FileCloseFailed → "file_close_failed" / 1003
• BadImageState → "bad_image_state" / 1004
• And others for each PAL failure mode

StatusDetailsExt

PAL implementations declare an associated type:

trait OtaPal {
    type StatusDetailsExt: StatusDetailsExt + Clone;
    fn status_details(&self) -> Self::StatusDetailsExt;
}

Custom fields are combined with base OTA fields via CombinedStatusDetails<E>, which flattens both into a single JSON object during serialization. Default impl for () provides zero-cost opt-out.

Integration Test Infrastructure (tests/common/aws_ota.rs)

• Loads management credentials from environment, assumes role into target IoT account
• Cleans up stale QUEUED/IN_PROGRESS jobs before each run
• Uploads firmware to S3, creates OTA update, polls until job execution appears
• Post-test: describe_job_execution() asserts cloud-side status and status details
• Guaranteed cleanup (cancel jobs + delete S3) even on test panic

Changelog

• Add OtaPalError variants with error_code() and as_reason_str() methods
• Add OtaStatusDetails struct with reason, error_code, and progress fields
• Add StatusDetailsExt trait and CombinedStatusDetails<E> for extensible job status details
• Add StatusDetailsExt associated type to OtaPal trait
• Preserve OtaPalError through error handling chain instead of discarding to Aborted(None)
• Include block progress count in Failed status details
• Add self-managed OTA integration test infrastructure with AWS credential management and cleanup