Check when telescope status in agc_exposure table is obtained

Telescope statuses stored in agc_exposure table look different from what is expected for the time when the image is taken.
The below is a copy from message on Slack, pfi-ics channel:
 
****
I have a question about the telescope status stored in {{agc_exposure }}table. We look for the best-focus position using AG; taking AG images at different hexapod z-position and getting the position where the spot size is the same between w/ and w/o glasses. Looking the data in opDB, it seems that telescope status is different from expected.
The below table is a example: we took three frames at each hexapod-z position, -0.3, -0.2, -0.1, 0, +0.1, +0.2, +0.3.  The spot size (the four columns on the right are average of spot size stored in agc_data table) seems to follow telescope movement; the size changes every three frames. On the other hand, m2pos3 (hexapod position) looks shifted by one... I guess the telescope status is for the last exposure.
Could you tell when you read telescope status, and write on opDB?

agc_exposre_id m2pos3 l_x l_y r_x r_y 
 191084 5.09 2.09 1.83 3.91 4.52
 191085 4.79 2.06 1.73 3.72 4.09
 191086 4.79 2.32 2.12 4.05 5.58
 191087 4.79 1.65 1.51 2.73 3.58
 191088 4.89 1.76 1.51 2.86 3.52
 191089 4.89 1.69 1.47 2.81 3.40
 191090 4.89 1.97 1.59 2.21 2.03
 191091 4.99 1.80 1.49 2.07 2.19
 191092 4.99 1.77 1.49 2.20 2.03
 191093 4.99 2.03 1.88 1.78 1.77
 191094 5.09 1.58 1.42 1.40 1.38
 191095 5.09 2.31 1.94 1.89 1.77
 191096 5.09 2.44 2.08 1.98 1.88
 191097 5.19 2.27 1.91 1.67 1.52
 191098 5.19 2.65 2.31 1.96 1.68
 191099 5.19 3.79 4.41 2.38 2.05
 191100 5.29 3.68 4.04 2.25 2.12
 191101 5.29 3.75 4.27 2.15 2.06
 191102 5.29 4.16 4.70 2.35 2.18
 191103 5.39 4.82 5.34 2.77 2.22
 191104 5.39 4.40 4.06 2.31 1.83