envoyproxy/envoy

Rethink ALL_CLUSTER_STATS

Open

#7,128 opened on May 31, 2019

View on GitHub
 (2 comments) (0 reactions) (0 assignees)C++ (5,373 forks)batch import
help wantedtech debt

Repository metrics

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

Description

We have been dumping a bunch of things in ALL_CLUSTER_STATS that are specific certain aspects of Envoy configuration and/or extensions. Given that these stats are the largest users of memory, and there is cognitive load for having stats that are not used, we should consider breaking this apart and allowing code/extensions to create sub-scopes in the cluster for specific things.

For example:

COUNTER(lb_local_cluster_not_ok)                                                                 \
  COUNTER(lb_recalculate_zone_structures)                                                          \
  COUNTER(lb_subsets_created)                                                                      \
  COUNTER(lb_subsets_fallback)                                                                     \
  COUNTER(lb_subsets_fallback_panic)                                                               \
  COUNTER(lb_subsets_removed)                                                                      \
  COUNTER(lb_subsets_selected)                                                                     \
  COUNTER(lb_zone_cluster_too_small)                                                               \
  COUNTER(lb_zone_no_capacity_left)                                                                \
  COUNTER(lb_zone_number_differs)                                                                  \
  COUNTER(lb_zone_routing_all_directly)                                                            \
  COUNTER(lb_zone_routing_cross_zone)                                                              \
  COUNTER(lb_zone_routing_sampled)                                                                 \

All of ^ could be created on demand by extensions/LBs that use them.

cc @jmarantz

Contributor guide

Research direction
Analyze the current ALL CLUSTER STATS macro and identify which stats are used by which extensions. Propose a design to allow sub scopes per extension, and create a proof of concept with one extension (e.g., load balancer) to validate the approach.
Tech stack
None
Domain
backendinfrastructure
Issue type
Refactor
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.
Over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
C++Envoy
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
30