Hello everybody out there using scope.
I've been using this project since early 2020 and genuinely liked it. It was one of the tools that made understanding a running Docker/Swarm setup at a glance actually pleasant. Once it stopped being maintained, I kept using it for a while, then eventually drifted away and stopped running it. What brought me back to it wasn't a sudden burst of free time. It was the emergence of agentic coding tools finally making a proper revival practical for one person to do in spare time: tracing a long-dormant bug across the codebase, swapping out a long-unbumped dependency, and actually validating the fix against a live cluster stopped being a multi-weekend undertaking.
So: we've been doing a modernization pass on Weave Scope for current Docker and Swarm (just a personal effort, not affiliated with Weaveworks or attempting to un-deprecate the project officially). My priorities so far have mostly been "does it still compile" and "does it still work against a current Engine" rather than anything more ambitious.
Fork: https://github.com/alfiyansys/scope
So far (as of 2026-08-05 18:00 +7):
- Bumped the Go toolchain, and replaced
fsouza/go-dockerclient with the official docker/docker/client SDK. The library itself is still maintained upstream, but this project had it pinned to a 2018-era release tracking a 2018-era Docker API version, so it was time to move to the SDK Docker maintains directly. Now running against Docker Engine 27.x/28.x.
- Found and fixed a long-standing bug while validating Swarm support:
report.Topology.AddNode() never stamped Node.Topology on dynamically-created nodes, so UnsafeRemovePartMergedNodes silently dropped every Swarm service node before it reached the UI. The "Services" tab has likely been empty for years for anyone running Swarm. Same bug, same fix, was also present in the ECS reporter.
- It's currently deployed as real infrastructure on my homelab's 4-node Swarm cluster, with a probe per node reporting to a central app, correctly showing cross-host containers and live cross-service traffic (proxy to backend, etc.) — screenshot below.
There's no roadmap and no commitment to maintain this long-term. It's one person's fork, published in case it's useful to anyone else still running Scope. If you hit something while using it, I'd be interested to hear about it, though I can't promise a fix.
Hello everybody out there using scope.
I've been using this project since early 2020 and genuinely liked it. It was one of the tools that made understanding a running Docker/Swarm setup at a glance actually pleasant. Once it stopped being maintained, I kept using it for a while, then eventually drifted away and stopped running it. What brought me back to it wasn't a sudden burst of free time. It was the emergence of agentic coding tools finally making a proper revival practical for one person to do in spare time: tracing a long-dormant bug across the codebase, swapping out a long-unbumped dependency, and actually validating the fix against a live cluster stopped being a multi-weekend undertaking.
So: we've been doing a modernization pass on Weave Scope for current Docker and Swarm (just a personal effort, not affiliated with Weaveworks or attempting to un-deprecate the project officially). My priorities so far have mostly been "does it still compile" and "does it still work against a current Engine" rather than anything more ambitious.
Fork: https://github.com/alfiyansys/scope
So far (as of 2026-08-05 18:00 +7):
fsouza/go-dockerclientwith the officialdocker/docker/clientSDK. The library itself is still maintained upstream, but this project had it pinned to a 2018-era release tracking a 2018-era Docker API version, so it was time to move to the SDK Docker maintains directly. Now running against Docker Engine 27.x/28.x.report.Topology.AddNode()never stampedNode.Topologyon dynamically-created nodes, soUnsafeRemovePartMergedNodessilently dropped every Swarm service node before it reached the UI. The "Services" tab has likely been empty for years for anyone running Swarm. Same bug, same fix, was also present in the ECS reporter.There's no roadmap and no commitment to maintain this long-term. It's one person's fork, published in case it's useful to anyone else still running Scope. If you hit something while using it, I'd be interested to hear about it, though I can't promise a fix.