10,000 sites migrated, here's what broke every time

Back in 2009 they don't make them like that anymore, but they also didn't have 47 different ways for a WordPress site to quietly rot. My agency just crossed 10,000 migrations and I started keeping a taxonomy because the patterns got depressing.

The big ones:

• DNS TTLs set to 86400 by hosts who should know better, clients wondering why "it's still pointing to the old server" 18 hours later
• Hardcoded IP addresses in wp-config.php from that one time in 2017 they "optimized" latency
• Cron jobs living in individual cPanel accounts nobody documented, sending emails to dead addresses until someone notices the bounce storm
• SSL certs with SANs for domains the client forgot they owned, auto-renewal fails, chain breaks
• Custom PHP extensions compiled against ancient libxml versions, new server has different ABI, white screen of death
• Email routing: the site moves, MX stays, but the contact form still tries local delivery because "it worked before"

The worst? The quiet breaks. Site loads, everything green in the monitor, but the webhook to their payment processor has been 404ing for six days because the endpoint was /old-server-ip/webhook.php.

Back in 2009 at least things failed loudly. They don't make them like that anymore.

What's your favorite migration horror story?

1) congratulations on the milestone
2) your taxonomy is incomplete
A) missing category: database encoding
I) latin1_swedish_ci to utf8mb4 migration
Ii) emoji in post titles becomes ?????
Iii) client notices only after new blog post goes live
B) missing category: serialized PHP arrays
I) search-replace on URL breaks string lengths
Ii) widgets silently vanish
Iii) no error in logs
3) emotions about this
A) frustration
I) high
B) admiration for your count
I) also high

Mas tipo the webhook thing happened to me last month!!! Client was losing and nobody knew why logs looked fine

DNS ⏰⏰⏰ is always the real villain tho