Add the missing failure_reason tag.

[wuhao5]

What was changed

When omitting the metric "WorkflowTaskExecutionFailureCounter", there are two other occasions that don't specify the additional "failure_reason" tag. See here that omits the same metric with additional tag.

Why?

When using prometheus as a client backend, it enforces the same metric be consistent with the same tags. If not, it will generate an error, and no metric is reported.

Checklist

1. Closes: simple change, no issue is created.

2. How was this tested: Implementing a new metrics.Handler that is backed by Prometheus, executing an intentional failing activity or workflow, no metrics for the errors are reported.

3. Any docs updates needed? N/A.

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}