Currently web scenarios stop when any step fails.  Many web scenarios log into a service, run some tests, and later log out.  In these situations, a failure means the logout never happens.  This causes issues for some services I monitor that have limits on the number of simultaneous sessions a user can have.  A temporary failure anywhere after logging in causes future scenarios to start failing at the login step.

I propose adding a flag to each web scenario step to indicate if it should always be run (even if a previous step fails).  The rest of the behavior can remain the same, with the failed step being reported even though a later step was run.

This is similar to ZBXNEXT-986, but not quite the same.  However, if stats are still gathered on the "always" steps, then this could work to satisfy that request too.