Is "Blocking" always client side?

So the HAR spec defines “Blocking” as:

Blocking - Time spent in a queue waiting for a network connection.

Does this mean that it is the time spent on the client waiting for a network connection, or time spent waiting for a server-side connection?

As far as HAR goes, yes - “blocked” is the time the client waited for an available connection to free up.

If there is a problem actually connecting to the server that will show up in the connect time.

If the server could accept the connection but can’t process the actual request then it will show up in the wait time.