e2e: increase client route propagation timeout to reduce flakiness

[juan-malbeclabs]

Summary of Changes

• Increase the client-side route propagation timeouts in TestE2E_MultiClientIBRLAllocatedIP from 60s to 90s, aligning them with the device-level iBGP checks that precede them in the propagation chain
• Fixes intermittent CI failures ("client1 should have route to client2", condition never satisfied) caused by the client checks expiring before routes finished propagating under load

Diff Breakdown

Category	Files	Lines (+/-)	Net
Tests	1	+2 / -2	+0

Pure test-timing change; no production code affected.

Key files (click to expand)

• e2e/multi_client_ibrl_allocated_ip_test.go — bump both client route-check timeouts from 60s to 90s to match the device-side iBGP propagation window

Testing Verification

• Routes reach a client only after its device learns them via iBGP (90s budget) and re-advertises them over the client BGP session; the client checks previously allowed only 60s, so a near-limit device propagation left insufficient time for the client install. Raising the client budget to 90s removes that race.
• Ran TestE2E_MultiClientIBRLAllocatedIP locally in a repeat loop to confirm it no longer flakes (first iteration passed in 666s; additional iterations in progress).