If you can’t see it, you can’t monitor it, and if you can’t monitor it, you can’t measure it, and if monitoring and measuring are not possible, then glitches will happen, people will complain, and your service credibility will be hit.

Deploying a network tap on a copper or fibre Ethernet connection will uncover low level link layer problems such as physical link interruptions, and also accurately mirror layer three, and above, communication protocol issues to your analysis tools as they happen. This is the ONLY reliable way of seeing the glitches in their native environment and it allows you to discover how issues occur and what they look like on the wire when they affect normal operational behaviour. This insight lets you set about resolving the root of the issues; other methods of observation simply highlight the sympathetic effects of these types of issue, so you are never really “squaring down” on the cause.
A span or mirror port cannot unearth the cause of ‘errored’ packets, in turn causing multiple retries that simply consume bandwidth and slow everything else held up behind them, or 1G links refusing to run faster than 100M with no-one knowing that it is simply due to a physical link error, amongst a variety of other issues; whereas properly placed network taps will unearth all of these low level types of issues.
With the other problems you have in your 100% “full-on” day, it’s nice to know that you might have 99 other problems, but that network glitch that was addressed with a network tap and just got sorted wasn’t one of them.
Apologies to Jay-Z!