0

I have a FeathersJS and Apollo graphql app that uses a Compose.io as the database. The portal servers have similar memory consumption but the data servers are very different (see screenshot below). A few weeks ago I increased the memory allocation because I was experiencing memory errors. The errors have stopped but one database is still at the memory limit and the other database is less than half.

Is this an indication of an underlying problem?

Screenshot of compose memory metrics

sagannotcarl
  • 354
  • 2
  • 13
  • How are these two data servers related and why would the database memory consumption be expected to be similar? What specific version of MongoDB and storage engine are these using ? It is quite possible that one data server instance requires access to a larger range of documents, has more active connections, or performs operations that require further allocation of memory. I think this is a better question to investigate with your database provider since they should have access to compare or provide metrics for both deployments. – Stennie Mar 19 '18 at 07:24
  • @Stennie: The two servers are shards and I have a single app that is accessing it. So my assumption was that they were handling the same data. The MongoDB version is 3.2.11. Thanks! – sagannotcarl Mar 19 '18 at 16:01

0 Answers0