Hi Grant. Looking into this. It looks like a few other instances were migrated as well for you

• prj=garden / svc=Postgres / env=production / deployId=b17857
• prj=css-paint / svc=Postgres / env=production / deployId=fae174
• prj=collage / svc=Postgres / env=production / deployId=660d76
• prj=quotes / svc=Postgres / env=production / deployId=17c45f
• prj=cascade / svc=Postgres / env=production / deployId=3e35e9
• prj=nest / svc=Postgres / env=production / deployId=94043d
• prj=db-feed / svc=Postgres / env=production / deployId=7138da

Do any of those have this issue? It looks like the migration completed successfully, and then, shortly after it started, there was a drop in the data

I'm looking into this with the team but, just wanted to see what the DB is used for an if there's a possibility some kinda of external job fired (like, what's upstream of this Postgres?)

Figured it out. Wow that's a doozie

So, when we create the migration, we don't actually move the old instance. We make a copy, streaming replicated, verify integrity, and then cut over.

It appears your volume was migrated during a high traffic time, which meant the migration timed out. Which, would have been fine, except it was retried. And when it retried, it tried to copy from new (empty) volume. We performed this migration 10-20k times yesterday, and it looks like < 10 instances right now were affected.

We've done the following:

- Fixed the bug

- Reset the migration, restoring your data

- Applied credits to your account for the inconvenience

Again, very sorry this happened. I can assure you the data is there, and that we take your trust in us, and your data, VERY seriously.

Best,

Jake Cooper

Founder @ Railway

So glad it's back! Thanks for figuring it out and the explanation.

Thanks for sticking with us through the migration!