Just a few points of verification, clarification, and tagging DetChar, letting them (especially the CW group) know this sweeping PCAL X calibration line has been turned on.

I attach a full and zoomed CAL-DELTAL spectrum against a PCALX spectrum (among the other usual things from the front wall's sensitivity FOM). 

Note that Pulsar injections are on-going, and it doesn't look like these features will interfere. This calibration line regime is sticking to the 2000 to 5000 Hz regions, and pulsar injections only go up to 1991.2 Hz (For a reminder of the full list of pulsar injection frequencies, see LHO aLOG 27642.

The starting frequency is 2001.3 Hz, and the starting excitation amplitude is 30000 [ct], and has been hard-coded to remain this amplitude for all frequencies. The excitation frequency can be tracked using EPICs via this channel: 
    H1:CAL-PCALX_PCALOSC1_OSC_FREQ
and you can see a fast-channel version of the requested output here:
    H1:CAL-PCALX_EXC_SUM_DQ

The guardian code for this node lives here 
/opt/rtcds/userapps/release/cal/h1/guardian/HIGH_FREQ_LINES.py

and is attached for ease of reference.

After some initial debugging, it guardian's log suggests that TJ left the frequency at 2001.3 starting at 2017-03-28 17:16:31 UTC, though it looks like the frequency has only been stable at 2001.3 since 19:49:00 UTC. TJ has the CHECK_IFO_STATUS state look whether the ISC_LOCK guardian's state is lower than 11, 
which means the excitation is killed if we're in
    - INITIAL ALIGNMENT
    - DOWN
    - IDLE
    - LOCKLOSS
    - LOCKLOSS_DRMI
    - INIT
but maybe we want to rethink this, because it seems to turn off the excitation at unnecessary times (see attached dataviewer trend, where the guardian has changed the frequency twice since TJ left it).