正在加载视频...

视频加载失败

Multigres is now in private alpha: high availability for Supabase Postgres. If your database fails, a standby copy takes over on its own, and data your app already saved is not lost. Organizations on paid plans can request access. Learn more →

17,463 次观看 • 4 天前 •via X (Twitter)

15 条评论

˗ˏˋ Jesse Hanley ˎˊ˗ 的头像
˗ˏˋ Jesse Hanley ˎˊ˗4 天前

.... so you just built the default @PlanetScale setup? Where is your response to Neki?

Luca Iaconelli 的头像
Luca Iaconelli4 天前

feednest and aevio both run on supabase postgres, so a db that fails over by itself without me waking up at 3am sounds pretty good to me

Pablo Manjarres 的头像
Pablo Manjarres4 天前

the standby waking up and taking the wheel is beautiful (my data staying saved should not feel like a feature but here we are)

Utkarsh Jha | building Piperoom 的头像
Utkarsh Jha | building Piperoom4 天前

Does failover carry logical replication slots? Postgres 17 can sync them to the standby (failover = true + sync_replication_slots). Before that, a Debezium or other CDC consumer lost its slot on failover, which usually meant a re-snapshot.

Ali 的头像
Ali4 天前

Can you announce things, when its already to use by public? It fustrates me when feature are announce but no where can be used e.g. OrieleDb

Dionis Ceban 的头像
Dionis Ceban4 天前

Failover only avoids data loss if the standby is synchronous, so synchronous_commit and replication lag are what I'd look at first

Code With Jerry | AI Tips 的头像
Code With Jerry | AI Tips4 天前

Automated failover for Supabase Postgres could make production apps more resilient by reducing downtime when a database instance fails.

MASA 的头像
MASA4 天前

How fast does the standby take over?

Hershal Rao 的头像
Hershal Rao4 天前

rip to everyone's excuses for missing production bugs

Abhishek kumar 的头像
Abhishek kumar3 天前

what would happen if if two nodes go down at once ? just curious 👀

Logs Portal 🚀 的头像
Logs Portal 🚀4 天前

Hey @grok can a ban account on supabase be recovered and how long

Anand Butani 的头像
Anand Butani4 天前

"data your app already saved is not lost" is the line that matters, since plain async failover quietly drops whatever hadn't replicated yet.

Goshen Web X 的头像
Goshen Web X4 天前

Automatic failover on Postgres without losing committed writes is the kind of thing you only appreciate after your first 3am outage. Curious how quickly the standby takes over in practice.

seridarivus 🇷🇴 的头像
seridarivus 🇷🇴4 天前

Automatic failover is the easy demo. The hard part is what the app sees during the switch: dropped connections, retries hitting the new primary, and long transactions that just vanish.

GreenLightMyApp 的头像
GreenLightMyApp4 天前

Hey @grok how many vibe coders developing new websites and apps use supabase?

相关视频