- Increase waits for system to settle
- Add polling for containers and edges
- Tabs formatting in setup.sh
- 300_internet_edge: Don't let curl fail in subshell (main source of flakyness)
- Make last param (count) optional for has_container (from Paul)
* Added helper for installing scope on gcloud
* Added topologies Pods and Pods-by-Service
* Uses k8s.io/kubernetes/pkg/client/cache for the client
* Filter kube-system nodes by default
* Only show the k8s topologies if we've received a non-empty k8s report
Also
- Add test for clustering without weave.
- Update test template for docker 1.8
- Prefetch images
- Stop all containers when test finishes.
- Add small sleep after weave launch, and supply all the hosts to ensure IPAM works correctly.
- Restructure main funcs for clean defer-stack-unwinds
- Fix Docker container to handle signals properly
- Introduce runsvinit for container init process
- Integration test
* always set updated topology object when received
* track whether route was set on initial load
* removed connection count from status bar (was not deemed important)
* fixed issue where topology option changes did not affect details pane
* only show filtered nodes when count > 0
* clear nodes graph when empty topology is loaded
* also prevent JS error if nodes are hovered over that should be gone
* fixed options sync issue between graph and status bar
* implemented topology options with immutable DS
- Process DNS resolution serially
- Process up to 10 HTTP GET (for app ID) concurrently
More than 10 concurrent GET requests will block on the semaphore. This
will cause the staticResolver.resolve method to block, which is probably
fine: it will just delay the next resolve loop, currently at 1m
intervals.
To make this a little bit more robust, I've also added a fastClient for
app ID resolution, with a timeout (total, including connect, request,
and response) of 5s.
