-
v1.2.22
StableSome checks failedSecret Scan / gitleaks (push) Successful in 22slint / lint (push) Successful in 45sGolang Tests / test-go (push) Successful in 1m2sRust Tests / test-rust (//rust/rperf-server:rperf, rust/rperf-server, bazel) (push) Successful in 4m9sRust Tests / test-rust (rust/consumers/zen, cargo) (push) Successful in 3m47sRust Tests / test-rust (rust/rperf-client, cargo) (push) Successful in 3m28sRust Tests / test-rust (rust/log-collector, cargo) (push) Successful in 3m50sRust Tests / test-rust (rust/srql, cargo) (push) Successful in 5m51sPublish Release Artifacts / publish (push) Successful in 24m21sRust Tests / test-rust (rust/trapd, cargo) (push) Has been cancelledImage Security Scan / image-security (push) Has been cancelledSource Security Scan / source-security (push) Has been cancelledreleased this
2026-04-22 03:40:52 +00:00 | 1674 commits to staging since this releaseServiceRadar v1.2.22
Docker Compose diagnostics and telemetry hardening, database workload isolation, and MTR automation fixes needed to keep fresh stacks responsive while running the full default control plane.
Whats New
1.2.22
- Fixed Docker Compose agent enrollment, MTR command dispatch/progress handling, and ICMP capability wiring so fresh compose installs can run diagnostics without empty agent selectors or stuck acknowledged jobs.
- Isolated control-plane database workloads and added runtime pool controls for
core-elx,web-ng, and agent gateway paths so long-running analytics, topology, NetFlow, trace summary, and maintenance work do not starve user-facing requests. - Reduced avoidable database pressure by skipping unchanged NetFlow reference refreshes, lowering heavy maintenance concurrency, adding query timeouts, and making NetFlow enrichment workloads feature-gated for compose environments.
- Made SNMP trap, flow, and BMP collectors opt-in compose profiles instead of default services, while preserving the Kubernetes/demo defaults and documenting how Docker users enable those collectors when needed.
- Hardened OTLP log/trace export and structured log handling so exporter failures and non-binary log messages do not flood compose logs or crash telemetry paths.
- Fixed MTR baseline cadence for bulk jobs by storing JSONB payload/context objects correctly and reading legacy double-encoded rows, preventing profiles configured for 300 seconds from dispatching every scheduler tick.
- Improved device details and analytics runtime behavior by deferring expensive fan-out work and clarifying migration/DB readiness messaging when pools are degraded for reasons other than schema migrations.
Downloads
-
Source code (ZIP)
1 download
-
Source code (TAR.GZ)
1 download