Shared Instances: Spikes causing context deadline exceeded

Hello there,

I started today monitoring the Dgraph instances (the $39.9/Month) and it seems there are frequent spikes happening throughout the day accross both eu-central-1 & us-east-1. I’m calling the endpoint with a simple Graphql query that returns 1 record, 1 field.

My users report random errors using the apps I’m running with a Dgraph Cloud backend.
From my understanding, DGraph Cloud returns “Context deadline exceeded” whenever the request takes more than 5 seconds.

eu-central-1 (spike today at 2PM CET-1)

us-east-1

Anything planned to fix those spikes? Or should we just host DGraph ourselves to avoid those issues?

Thanks

2 Likes

Updated charts for this afternoon:

Its not getting any better

2 Likes