The tunnels were back.
I sent the update after another troubleshooting pass on Peek's mobile-network integration. The firewall showed both secure connections up again. Once the carrier updated the remaining endpoint, we could test traffic across the full path.
This project had followed me out of the company.
I had worked on it while I was Peek's Director of Technical Operations. After I left for Reality Check, the team asked for help finishing it. Authentication traffic was not reaching the right place, the network had been divided in a way that complicated the route, and several organizations held different pieces of the system.
I wanted to help. I had left because of the company's financing problems, not because I stopped believing in Peek.
There was also a practical business conversation around it. Peek did not have the data-center setup I needed for a separate hosting project, so the broader arrangement we had discussed did not make sense. I proposed a modest support retainer instead, including help with the network and availability when something broke.
That let both sides be honest. Peek needed continuity. I needed to put a boundary around work that could otherwise expand indefinitely.
The final troubleshooting involved vendor documentation, new network information, and another round of checks. I asked a senior contact at Juniper whether the high-availability design would work through the switching layer. His answer was wonderfully short: it works.
Then the firewall showed the tunnels again.
The message did not claim the entire project was finished. One piece on the carrier side still had to change before the end-to-end tests.
But the secure path was up. After months of handoffs and partial fixes, that was real progress.