features/monitoring

[giscus[bot]]

features/monitoring

GraphQL Hive Documentation

https://docs.graphql-hive.com/features/monitoring

[yuhanz]

What is the overhead of enabling operation monitoring? Are the access logs transmitted to GraphQL-hive server through HTTP calls? I imagine the requests may be sampled or batched, and wonder if this feature increase memory usage, network bandwidth, connections, etc.

[dotansimha]

Hi @yuhanz !
You are correct, we are batching and optimizing the storage, and then shipping it together over HTTP. You can find the code here: https://github.com/kamilkisiela/graphql-hive/tree/main/packages/libraries/client

Memory increasing depends on your configuration, you can change the max config field and change the sampling count, and adjust the data stored in memory. We found the default configuration as working pretty well for small and large projects.

[pierrebiver]

I love your product :) and I know doc can be hard to write, but I actually stumble upon the router alternative by hazard when trying to self host it. I did not know it was a possibility (I went with Apollo server) it would be nice to make it somewhat more visible on the self hosting part maybe ?

[pierrebiver]

I meant like having in the self hosting and get started section a mention to the possibilities on how to implement the client ? I just followed the video tutorial.

[pierrebiver]

also, what kind of permission the token requires in the example above ?