sendtracker

README

Overview

This pod is used to measure unreceived (but accepted) message sends.

Details

This pod is designed to be both an event sender and an event receiver. It sends a series of events through a single threaded sender, keeping track of the response code for each send. The receiver maintains a list of received events. When stopped, the receiver then waits several minutes for all events to be delivered.

The ranges of events successfully sent and received are then compared, with any mismatches seen. It is expected that no events will have been successfully sent, but not received. These are returned as failures. It is possible (on the other hand) that an event may not have been successfully sent, but will be received. For instance, if the ingress of a broker is killed after it enqueues onto pubsub, but before the handler returns). These are reported as warnings. The number of duplicate receive events seen is also reported (no duplicate sends are made, but reliable forwarders like the broker may forward some events more than once if they are unsure that the previous forward worked)

The yaml's in this directory hard code it to send to the knative GCP broker for the default namespace and to receive events via a trigger. As a prerequisite, the broker must already be configured for this namespace. It must be run as a deployment with a single pod. They can be deployed with ko apply -f .

This pod then provides a very basic 5 call http "api" to control data taking. This API can be reached on the pods port 8070 either via kubectl port forwarding or via a curl pod run inside the cluster.

Control

• curl sendtracker:8070/ready returns "true" once the sender has successfully sent a startup message to the receiver.

• curl sendtracker:8070/start starts sending events to the receiver. After this is run, the user should then do whatever perturbations are desired to test (upgrades, failure injection, etc)

• curl sendtracker:8070/stop stops sending events to the receiver, but waits a configurable time for events to be delivered

• curl sendtracker:8070/resultsready returns "true" once the receiver has finished listening, and results are ready.

• curl sendtracker:8070/results returns a human readable summary of the run. The first line is "success" if no errors were seen and "failure" if errors were seen. The following lines are human readable info about errors, warnings, all send results, all events received, and the number of duplicate events seen.

To run again, delete the pod and let the deployment restart it.

Tuning

There are 3 parameters that can be tuned for the deployment.

• K_SINK - The target for the events. This target should eventually feed these events back to the pod.
• DELAY_MS - The time to wait (in milliseconds) between event sends
• POST_STOP_SECS - The time to wait (in seconds) between the user request to stop and the receiver stopping to record received events.

Documentation

Overview

Copyright 2020 Google LLC

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.