How do you move an undocumented legacy app to the cloud without breaking everything?

November 14, 2026 4 answers
Asked by p407!stencil_8

There's an old web application on failing on-premise hardware, with little documentation and a tight deadline to move it. How would you approach the migration when the dependencies and database risks aren't fully understood?

Answers

u26!stencil_935

Given the failing hardware and deadline, I'd look at lift and shift first. Get the existing application running on a cloud instance with as few changes as possible, then tackle modernization once it has a safer home.

That doesn't mean blindly copying it over. You still need to identify what it talks to, how data gets backed up, and whether you can restore it. An old runtime may also need extra isolation while you plan an upgrade.

w41~notch_709:

A temporary move can buy time, but make sure the temporary setup has an owner and a plan for dealing with unsupported software.

Reply to this answer
p706!wordmask_4

Start with the database, not containers. Verify that backups exist and test a restore somewhere separate before making migration changes. Work out how you'll handle writes during the cutover and what would make you stop or roll back.

Then inventory the application as it actually runs: services, scheduled jobs, environment variables, file storage, network connections and anything else it depends on. The code alone may not tell you what the server is doing.

p407!stencil_8:

Testing the restore separately makes sense. A backup isn't much comfort if nobody knows whether it works.

Reply to this answer
almost.unindexed_4

Bring in migration help if the database or infrastructure is beyond the team's comfort zone, but keep someone internal involved. Ask for a clear scope, a migration plan and documentation of what they discover, rather than paying for a handoff nobody can maintain.

Reply to this answer
Dremby.4

Management needs to hear that moving to the cloud and fixing years of tech debt are different jobs. The immediate goal may be getting off failing hardware, not rewriting the app by the deadline.

Make the risks and unknowns visible, including what happens if the old server fails before the move. That gives everyone a clearer basis for deciding what can wait and what needs attention now.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions