Shifts will either be eight hours or 12hr shifts.
The pressure arises from working in what is a cerebral job while all around you, you'll have TRCs, ICs and supports all yelling across the room especially during an incident or emergency. Compound to this, you'll often be the only TDA covering a very large or very busy patch so if you fall behind, you're screwed.
Furthermore, you're expected to resolve/code a delay roughly one every two minutes which frankly is bonkers considering the depth of information you have to trawl. Easily done if you can quickly see on TWM that your train was held up behind another train, and that the other train's holdup is already coded / attributed, or that if your region has access to the TOC concerned's train manager/guard reports, they've left a note explaining why they were late ("4min late off X due to wheelchair and disruptive passengers"). However not so easy if there's nothing as then you will need to watch the movement replay on TVM, go through incident logs, then start flicking through multiple pages of TRUST (which is a COBOL based mainframe access) which is time consuming.
Or then your specific delay is say 5mins, you look into it, and it was knocked off by a train running 2mins late, and that train was knocked out by another train only two mins late, and then that third train was 2 late because a fourth train delayed it by 2-3mins. Because then you have to create cases for each of these delays, classified as "sub thresholds" and investigate each of =those= delays so that you can create the case for the last one, then link the one after that, link the third one to that, so in the end you can then link your specific delay to all of that. You can end up spending 30mins plus to account for 3mins of actual presented delay. (and then your manager will possibly be stood there, tapping their foot, demanding why your output rate was slow.)
In a major incident, the ripple effect will be horrific. Train after train will be delayed and each one will present on screen rapidly. You can have a backlog of over 100 delays all requiring the same meticulousness and needing to be done within a certain time but frankly, as someone who has done it for a good while, you will almost never have it wrapped up until long after.
There is no set progression from TDA; however you may be able to ask for days where you can shadow a signalling team, a manager or a TRC/IC to gain insights into those roles as often in our ROC TDAs go onto TRC and IC roles, or management roles.
TDA is a job I would consider carefully. However, when you apply, you sit the same psychometrics as you would to go into a signalling role (multi tasking and so on) so if you are successful you could use it as a stepping stone, pass out from your probation and then consider your next steps. TDA is on the signalling grades in terms of Grade, Ts and Cs (though beware: TDA is not a safety critical role, so you are not safeguarded in the same way as actual signallers are. It is advertised as safety critical, but you are not actually under the same means.)