Long(ish) network latency between Railway and Planetscale
ekosz
PROOP

3 months ago

Hi there,

We're trying to use a Postgres Planetscale instance in combination with Railway, but we're seeing high than expected network latency between the two services. Looking at our logs most of our queries in Planetscale are taking less than 1ms (there's very little data at the moment). But in Railway I cannot find a SQL statement that takes less than 40ms to complete.

For context, our application is deployed in us-west-2 and and also picked that region for Planetscale. The DB connection URL we're using is us-west-3.pg.connect.db. When I ssh into the instance and run netcat, I see its about a 13-17ms ping time. I've also tried installing pscale and running it's ping command and it confirms the latency (see below).

My questions are:
1. With a 13-17ms ping time does a 40ms floor for SQL statements make sense?

  1. Is a 13-17ms ping time expected when the two services are in the same region?

Thank you!

=====

root@7435190a9d89:/app# pscale ping
  NAME (19)                     LATENCY   ENDPOINT                                         TYPE
 ----------------------------- --------- ------------------------------------------------ -----------
  AWS us-west-2                 13.6ms    us-west.connect.psdb.cloud                       direct
  AWS us-west-2                 14.4ms    aws.connect.psdb.cloud                           optimized
  GCP us-central1               36ms      gcp.connect.psdb.cloud                           optimized
  GCP us-central1               37.1ms    gcp-us-central1.connect.psdb.cloud               direct
  AWS us-east-2                 50.7ms    aws-us-east-2.connect.psdb.cloud                 direct
  GCP us-east4                  55.8ms    gcp-us-east4.connect.psdb.cloud                  direct
  GCP northamerica-northeast1   61.3ms    gcp-northamerica-northeast1.connect.psdb.cloud   direct
  AWS us-east-1                 62.2ms    us-east.connect.psdb.cloud                       direct
  AWS ca-central-1              75.4ms    aws-ca-central-1.connect.psdb.cloud              direct
  AWS ap-northeast-1            93.2ms    ap-northeast.connect.psdb.cloud                  direct
  GCP asia-northeast3           118.5ms   gcp-asia-northeast3.connect.psdb.cloud           direct
  AWS eu-west-2                 129.8ms   aws-eu-west-2.connect.psdb.cloud                 direct
  GCP europe-west1              134.9ms   gcp-europe-west1.connect.psdb.cloud              direct
  AWS eu-west-1                 136.8ms   eu-west.connect.psdb.cloud                       direct
  AWS eu-central-1              140.5ms   eu-central.connect.psdb.cloud                    direct
  AWS ap-southeast-2            153.5ms   aws-ap-southeast-2.connect.psdb.cloud            direct
  AWS ap-southeast-1            165.1ms   ap-southeast.connect.psdb.cloud                  direct
  AWS sa-east-1                 173.4ms   aws-sa-east-1.connect.psdb.cloud                 direct
  AWS ap-south-1                219ms     ap-south.connect.psdb.cloud                      direct
Solved
+11

1 Replies

brody
EMPLOYEE

3 months ago

The latency you're seeing makes sense given the setup. We don't run on AWS or GCP, so even though both services are nominally in "us-west-2", they're actually in different data centers communicating over the public internet. That 13-17ms ping time is the baseline network round-trip, and when you factor in TLS handshake overhead, connection establishment, and the actual query execution, 40ms+ for a complete SQL statement is reasonable.

If you're looking to reduce this latency, you might consider hosting your database on Railway itself using our Postgres offering, that would allow you to use private networking between your app and database, which eliminates the public internet hop entirely and typically brings query times down significantly to 1-2ms in our testing.

Status changed to Awaiting User Response Railway • 3 months ago

Railway
BOT

3 months ago

This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!

Status changed to Solved Railway • 3 months ago

Loading...