> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/continuous-delivery/troubleshooting-and-resources/armory/general/validator-validate-halvalidator-detected-a-fatal-error.md).

# Validator \*validate.halValidator detected a fatal error

### Issue <a href="#issue" id="issue"></a>

After upgrading Operator to version 1.4.1+ and Halyard to 1.12.0+, the following errors are seen on the Operator logs. It should be noted there were no changes to the Canary configs before and after the upgrade.

```
{"level":"info","ts":.....,"logger":"spinvalidate","msg":"Validator *validate.halValidator detected a fatal error"}
{"level":"error","ts":...,"logger":"spinvalidate","msg":"\nSpinnakerService validation failed:\njson: cannot unmarshal number into Go struct field PrometheusCanaryServiceIntegration.metadataCachingIntervalMS of type bool\n
```

Despite the removal of the field `metadataCachingIntervalMS` from the canary configs, the error remains in the Operator logs.

### Cause <a href="#cause" id="cause"></a>

Armory Engineering is aware of the issue. This has been identified to be an issue while parsing the configuration for Prometheus within Operator. It is a bug introduced when migrating the validators to operator.