jaegertracing/jaeger

gRPC plugin framework should be able to recover from panics

Open

#1742 opened on Aug 19, 2019

View on GitHub
 (7 comments) (0 reactions) (0 assignees)Go (18,974 stars) (2,326 forks)batch import
area/storagehelp wanted

Description

Requirement - what kind of business use case are you trying to solve?

We are implementing a custom gRPC-based storage plugin as per this doc.

Problem - what in Jaeger blocks you from solving the requirement?

There are two related problems:

  • When gRPC plugin panics the reason for panic is not shown in Jaeger logs, making the cause very hard to identify (related to: #1529)
  • When plugin panics it is not restarted, so Jaeger enters an unusable state

Impact:

  • Makes it very hard to make any plugin prod-ready
  • Makes developer experience of writing a plugin very frustrating
  • Plugin code becomes littered with defer/recover to prevent it from fully crashing and display helpful debug info

Proposal - what do you suggest to solve the problem or improve the existing situation?

Reason and a stacktrace should be already dumped into stderr of a plugin process at the time of panic, so Jaeger should be able to capture and log it.

Crashed plugins should be restarted.

Any open questions to address

Contributor guide

Tech stack
gogrpc
Domain
backendobservability
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
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.
clear
Prerequisites
Go programminggRPC basicsJaeger plugin system
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50
Research direction
Investigate the plugin lifecycle in plugin/storage/grpc/README.md and related code. The panic recovery should capture stderr output from the plugin process and log it. Also consider restarting the plugin process after a crash. Check issue #1529 for related logging improvements. Look at the existing plugin management code in the repository to understand the current startup and health check mechanisms.