envoyproxy/envoy

Envoy and control-plane xDS over ADS protocol deadlock when filtering stats

Open

#9,856 opened on Jan 28, 2020

View on GitHub
 (12 comments) (0 reactions) (0 assignees)C++ (5,373 forks)batch import
area/statsarea/xdsbughelp wanted

Repository metrics

Stars
 (27,997 stars)
PR merge metrics
 (Avg merge 8d) (378 merged PRs in 30d)

Description

Description: When using a stats_matcher inclusion_list, Envoy using the ADS variant of the xDS protocol will fail to receive cluster updates from the control-plane. After receiving an initial Cluster list, Envoy will no longer receive cluster updates. When paired with envoyproxy/go-control-plane in ADS mode, this can cause EDS updates to stall if the cluster list is later expanded.

Repro steps: A docker-compose reproduction of the issue with Envoy 1.12.2 and 1.13.0

Relevant Links:

Contributor guide

Research direction
Investigate the interaction between stats matcher inclusion list and the ADS xDS protocol in Envoy. Look at how the control plane (go control plane) handles incremental updates when stats filtering is active. Focus on the code path where cluster updates are missed after initial delivery.
Tech stack
cppgo
Domain
infrastructurebackend
Issue type
Bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
C++GoDocker
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40