sdk
sdk copied to clipboard
networkservicemesh
## Description We need to capture the latency/throughput of the current nsmgr. For this can be written benchmark tests. Attach results to this issue.
## Steps to reproduce 1. create nse that opens some resource on request by connId or use ipam server. 2. add to forwarder logic that on backward something happens and...
If connection is healed, new (or old) endpoint could assign IP addresses different from initial ones. Client does not know anything until it send refresh request.
# Intro Since we are running our unit tests with `--race` flag it dramatically (>10 times) slows down test execution. The cause for #755 was a GC STW overlapping with...
## Steps to reproduce 1. Create NSC that sets the kernel local mechanism in the request.Connection.Mechanism. 2. Deploy nsmgr, forwarder, nsc, nse Actual: NSC cannot connect to NSE. Expected: NSC...
Currently, we have a cover that selecting local and remote mechanisms are working via integration testing. We can cover this functionality via sandbox tests and 1. Verify do we really...
## Description Some of the errors that can be returned from the network service are errors from other tools/pkgs for example url.Parse() and so on. These errors are usually have...
## Motivation nsmgr could detect endpoints/forwarders death momentarily if we add functionality for monitoring netns. Currently, we are using [netns monitor](https://github.com/networkservicemesh/networkservicemesh/blob/027edac3488b4c511df465c36b62889586d81820/forwarder/pkg/common/monitor_netnsinode_server.go#L49) in monorepo in healing scenarious. ## Implementation details 1....
Network visualization is a powerful tool for users. It helps troubleshooting, clarify and define the relationships between network nodes and the links between them. The main visualization tool in monorepo...
Currently, Jaeger tracing is enabled on a per-application basis. (see #556) It would be convenient to enable it on a per-request basis. E.g. suppose Jaeger tracing is enabled globally on...
