What should I do if Neon is down for me?

Check your connection limit and pool settings first, since exhausting connections looks exactly like the database being down. Retry with a longer timeout, because a serverless or scale-to-zero instance may simply be waking up. Confirm your credentials and IP allowlist have not changed, as an auth or firewall denial reads as a failed query. Query from another network or client to establish whether the cluster or your route to it is the problem.

Databases

Is Neon down right now?

No, it’s up.

Neon appears to be up and running normally.

checked 2026-09-19 02:24 UTC

responded in 73ms

HTTP 200

Over the last 7 days, Neon has answered our automated checks with 100.00% availability and a typical response time of ~90 ms. No outages have been recorded in the last 90 days.

Neon is serverless Postgres with storage and compute separated, giving instant branching and scale-to-zero. Cold starts are part of normal operation, which makes diagnosis harder: a genuine incident and an idle compute waking up can look similar, and the control plane can fail while existing connections continue.

Is Neon down for everyone or just you?

Our automated checks are reaching Neon normally right now, and it answered our last one in 73ms. That makes any trouble you are hitting more likely to be local to your connection, account or network than a full outage, which usually means you can fix it from your end.

Still not working for you?Try the fixes

Neon response time (last 24h)

Last 24h

peak 577 ms

Neon incident history (last 90 days)

JulAugSep

up
slow
down
no reading

No incidents in the last 90 days, from our checks or Neon’s own status page.

Neon uptime

100.00%
7 days: 100.00%
30 days: 100.00%
90 days

Measured since 2026-08-12. Windows reaching further back are averaged over the 39 days we watched, not the full period. Counts our own checks. Days Neon reported an incident are coloured in the timeline above but do not change this figure.

What to do if Neon is down for just you

Check your connection limit and pool settings first, since exhausting connections looks exactly like the database being down.
Retry with a longer timeout, because a serverless or scale-to-zero instance may simply be waking up.
Confirm your credentials and IP allowlist have not changed, as an auth or firewall denial reads as a failed query.
Query from another network or client to establish whether the cluster or your route to it is the problem.

Frequently asked questions

Neon appears to be up and running normally. Last checked 2026-09-19 02:24 UTC.

Related services

ClickHouse Cloud CockroachDB Cloud Confluent Cloud Firebase MongoDB Atlas Pinecone