Get rid of noisy task adding failure log in matching service

[taylanisikdemir]

What changed?
Failed to add task error log for remote sync match failed scenario seems to be too noisy (96% of matching logs). We log it for multiple cases such as

• task forwarded from child partition but domain is not active in current cluster (ref)
• task forwarded from child partition but couldn't be sync matched (ref)

There's already a metric cadence_errors_remote_syncmatch_failed_per_tl counting this specific case and it has the same breakdown as the log does

• domain
• tasklist name
• operation (adddecision vs addactivity task)

So the change is to remove this log line. Other sub categories of Failed to add task error log will still be logged by the matching context's handleErr as an Uncategorized error. This seems fine to me but open for feedback here. We can keep those other sub types around and just get rid of remote sync match failed because it's a known type w/ its own counter.

Misc Change
Only signal taskReader if there's no error and no sync match found. Previously it was signaling even for sync matches but that looked unnecessary as nothing is written to persistence for sync match AFAICT. This can help reduce the unnecessary DB calls during sync matches.

Why?
Since this log entry adds no extra visibility and floods the logs, there's no point of keeping it.

How did you test it?
make test

Potential risks
Not signaling task reader for sync matches can have side effects I am not able to see. Looking forward to review feedback on this.

[CLAassistant]

All committers have signed the CLA.

[davidporter-id-au]

For higher-risk changes I'd suggest doing some integration-deployments, will discuss offline. otherwise LGTM