client/daemon: initial route liveness probing

[snormore]

Summary of Changes

• Adds support for Route Liveness Probing in the doublezerod client daemon, opt-in via the --route-probing-enable flag.
• Introduces a new probing subsystem integrated at the BGP plugin level, defined per service type (IBRL, IBRL with allocated IP, multicast).
  • Currently, probing is enabled only for the IBRL (without allocated IP) service, where fallback reachability over the public internet path is built-in.
• Performs ICMP-based reachability checks for all BGP-learned routes:
  • Routes in UP state are installed in the kernel routing table.
  • Routes in DOWN state are withdrawn from the kernel routing table until recovery.
• Implements a hysteresis-based liveness policy, requiring configurable consecutive successes or failures before transitioning between UP and DOWN.
  • The policy layer is pluggable, allowing future liveness evaluation strategies.
• Adds a user-space ICMP listener that receives echo replies over the doublezero0 interface even when the corresponding route is not installed in the kernel table.
  • Without this, the kernel ICMP responder would send replies over the internet rather than through doublezero0, preventing successful validation of otherwise reachable peers.
• Closes client/probing: initial daemon integration #1961
• Context: RFC: Client Route Liveness Probing #2018
• NOTE: Prometheus metrics will be added in a follow-up PR: client/probing: prometheus metrics #1964

Testing Verification

• Unit test coverage for probing manager, worker, and liveness tracker/policy.
• End-to-end tests blocked until Fix client to client connectivity testing in containerized e2e environment #2016
• Verified in local containerized devnet before intra-metro routing was disabled.
• Deployed and tested successfully in DZ devnet.

[Copilot]

Pull Request Overview

Adds support for Route Liveness Probing in the doublezerod client daemon, opt-in via the --route-probing-enable flag. This introduces a new probing subsystem integrated at the BGP plugin level that performs ICMP-based reachability checks for BGP-learned routes with hysteresis-based liveness policy.

• Implements route liveness probing for IBRL service with ICMP echo requests and hysteresis-based state transitions
• Adds user-space ICMP listener for probe responses over the doublezero0 interface
• Integrates probing system into BGP plugin with configurable route manager abstraction

Reviewed Changes

Copilot reviewed 44 out of 45 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tools/uping/pkg/uping/listener.go	Changes logging from Info to Debug level for listener events
tools/uping/cmd/uping-send/main.go	Adds Linux-only build constraint
tools/uping/cmd/uping-recv/main.go	Adds Linux-only build constraint
tools/twamp/pkg/light/stub_fallback.go	Removes deprecated build constraint syntax
go.mod	Updates prometheus-community/pro-bing dependency version
e2e/multi_client_test.go	Disables route probing in E2E tests until containerized connectivity is resolved
e2e/internal/rpc/agent_test.go	Adds binary availability check for doublezero
e2e/internal/devnet/device.go	Increases device container memory allocation
e2e/internal/devnet/cmd/add-client.go	Adds route probing enable flag
e2e/internal/devnet/client.go	Adds route probing configuration support
e2e/ibrl_test.go	Removes unused parameter from test helper function
e2e/docker/client/entrypoint.sh	Adds client flags environment variable support
e2e/docker/client/Dockerfile	Adds uping tools and additional packages
e2e/docker/base.dockerfile	Builds uping tools in base image
client/doublezerod/internal/services/services_test.go	Updates service creation calls for new route manager abstraction
client/doublezerod/internal/services/multicast.go	Adds route manager integration
client/doublezerod/internal/services/ibrl.go	Adds route manager integration and fixes error message
client/doublezerod/internal/services/edgefiltering.go	Adds route manager integration
client/doublezerod/internal/services/base.go	Adds NewRouteManagerFunc type definition
client/doublezerod/internal/runtime/run_test.go	Updates tests for new runtime Run signature
client/doublezerod/internal/runtime/run.go	Refactors to accept NetlinkManager parameter
client/doublezerod/internal/routing/routes.go	Adds nil safety checks to Route.String() method
client/doublezerod/internal/routing/netlink.go	Improves error handling for route deletion
client/doublezerod/internal/routing/errors.go	Adds ErrRouteNotFound error
client/doublezerod/internal/probing/	Complete probing subsystem implementation with worker, manager, liveness tracking, and configuration
client/doublezerod/internal/manager/services.go	Adds service creation functions with route manager abstraction
client/doublezerod/internal/manager/manager.go	Updates manager to use pluggable service creation
client/doublezerod/internal/manager/http_test.go	Updates tests for new manager constructor
client/doublezerod/internal/bgp/plugin.go	Replaces RouteReaderWriter with RouteManager interface
client/doublezerod/internal/bgp/bgp_test.go	Updates tests for new route manager interface
client/doublezerod/internal/bgp/bgp.go	Updates BGP server for route manager integration
client/doublezerod/cmd/doublezerod/main.go	Adds route probing command-line flags and configuration
CHANGELOG.md	Documents new route liveness probing feature

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.