CI timeout increase (and other improvements)

This MR does the following:

• increase how much time we spare to ensure debug-on-exit can run (from 200s to 450s) -- this should help avoid runs where the gitlab job timeout triggered during debug-on-exit information collection or node log collection
• in CI, output debug information for sylvactl --timeout computation
• CI: increase gitlab job timeout for deployment jobs

🍥 All of this was already tested in the context of !3090 (closed) (later wrapped into !2961 (merged)), because of the extended times that the upgrade from 1.1.1 takes.

I expect the third change to improve some nightly pipelines failures where timeout occur when only units at the end of the dependency chain remain pending reconciliation.