I am hesitant to do so because it has according to the seller "multiple sealed calibration stickers" and opening the case to change the caps would ruin that but I guess it is worth it in order not to have leaky caps
If all of the following apply, then don't open it. In general, though, it's better to have a working DMM than a dead one that's "in calibration."
1. The recalibration date hasn't yet passed.
2. You have the calibration certificate from the lab that calibrated it and its dates match the cal sticker.
3. The work that you do is accountable for its accuracy and you must be able to prove that your measurements and the equipment used to make those measurements follow specific standards.
For typical uses, regular traceable calibration is usually unnecessary. I'm part of the US Calibration Club on the forum and on the last go-round of the Geller SVR-T 10V reference we were trying out, my meters measured as shown below. Looks pretty stable to me.
Bench DMMAgilent 34410A: 10.00010 (Cal 8/2009)
Agilent 34401A: 10.00026 (Cal 7/2003)
Keithley 2700: 10.00007 (Cal unknown)
Keithley 196: 10.00035 (Cal unknown)
Fluke 8100A: 10.002 (Cal unknown)
Handheld DMMAgilent U1272A: 9.978 (Cal 6/2012)
Keysight U1282A: 9.998 (Cal 12/2016)
Keysight U1252B: 10.000 (Cal 10/2016)
Greenlee DM-820A: 10.00 (Cal unknown)
Note that although I don't know the exact age of my Fluke 8100A, they were made from 1969 through 1972 (based on the manuals that I have). So, if it had never been calibrated since manufacture, that's quite a long time to hold its calibration and it's pretty much bang on 10V.