Gloo supports a variety of configurations and options that can be accessed via Kubernetes resources or the Gloo helm chart.
- Usage Statistics
- Exposing Gloo with NodePort
- Configuring Function Discovery
How to monitor and trace within your Gloo setup.
Usage Statistics Gloo will periodically collect usage information from running instances. The details of this collection can be found here. The checks are performed on each invocation of glootcl after a given interval. This interval can be changed by setting the env var CHECKPOINT_TIMEOUT in milliseconds. This information icludes the following: Glooctl version Architecture Operating System To turn off these checks simply set CHECKPOINT_DISABLE=1
Motivation By default, microservices deployed in Kubernetes have an internal flat network that is not accessible from the outside of the cluster. This is true even if you use Kubernetes on a public cloud (like Amazon AWS or Google Cloud). A NodePort service is a way to make Kubernetes services available from outside the cluster (and potentially allow access from the internet) by opening ports on all of the nodes in the cluster and allowing traffic to go directly to the pods running within the cluster.
Motivation Gloo’s Function Discovery Service (FDS) attempts to poll service endpoints for: A path serving a Swagger Document. gRPC Services with gRPC Reflection enabled. This means that the Gloo discovery pod/binary will make HTTP requests to all services known to Gloo. This behavior causes increased network traffic and may be undesirable if it causes unexpected behavior or logs to appear in the services Gloo is attempting to poll.