docs: address issue #24139

[github-actions]

Summary

Added comprehensive documentation of Docker Prometheus metrics including descriptions, types, and usage examples.

Changes

• Added "Available metrics" section to Prometheus documentation
• Documented metric types (Counter, Gauge, Histogram) with explanations
• Explained histogram suffixes (_bucket, _count, _sum) with practical examples
• Added table of 11 engine_daemon_* metrics with descriptions
• Added table of 15 swarm_* metrics with descriptions
• Organized metrics by category (Engine vs Swarm) for easy reference

Fixes #24139
Fixes #19759

---

🤖 Generated with cagent

[netlify]

✅ Deploy Preview for docsdocker ready!

Name	Link
🔨 Latest commit	77294e1
🔍 Latest deploy log	https://app.netlify.com/projects/docsdocker/deploys/6992fa58c2dbcc0008914ed1
😎 Deploy Preview	https://deploy-preview-24148--docsdocker.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[usha-mandya]

@thaJeztah Could you PTAL?

[dvdksn]

I asked opus to review this PR against moby source code and here's what it had to say:

---

Thanks for working on this — documenting Prometheus metrics is valuable. I cross-referenced the metrics descriptions against the moby/moby source code and found several inaccuracies that need to be fixed before this can be merged.

Incorrect label references (engine metrics)

• engine_daemon_container_actions_seconds: The description says "start, stop, create, etc." but the actual labels in the source are: start, changes, commit, create, delete. There is no stop label. (source: daemon/internal/metrics/metrics.go:56-64)
• engine_daemon_network_actions_seconds: The description says "create, connect, disconnect, etc." but the actual labels used in the source are: update, allocate, connect, release. There is no create or disconnect label. (source: daemon/container_operations.go)
• engine_daemon_events_total: The description says "Labels indicate the event action and type" but this is a plain counter with no labels. It's defined as metricsNS.NewCounter("events", "The number of events logged"). (source: daemon/internal/metrics/metrics.go:49)

Inaccurate descriptions (swarm metrics)

• swarm_dispatcher_scheduling_delay_seconds: PR says "Time from task creation to scheduling decision." The source help text says: "Scheduling delay is the time a task takes to go from NEW to RUNNING state." — this measures the full NEW→RUNNING duration, not just scheduling. (source: vendor/.../dispatcher/dispatcher.go:74-75)
• swarm_raft_snapshot_latency_seconds: PR says "Time taken to create and restore Raft snapshots." The source says "Raft snapshot create latency." — "restore" is not part of this metric. (source: vendor/.../raft/storage.go:26-27)
• swarm_raft_transaction_latency_seconds: PR says "Time taken to commit Raft transactions. Measures consensus performance." The source just says "Raft transaction latency." The additions ("commit", "consensus performance") are not grounded in the source. (source: vendor/.../raft/raft.go:207)
• swarm_store_lookup_latency_seconds: PR says "Time taken for lookup operations in the swarm store." The source help text is "Raft store read latency." (source: vendor/.../store/memory.go:105-106)
• swarm_store_memory_store_lock_duration_seconds: PR says "Duration of lock acquisitions in the memory store." The source says "Duration for which the raft memory store lock was held." These are different things — acquisition time vs. hold duration. (source: vendor/.../store/memory.go:109-110)

Missing metrics

The following metrics exist in the source but are not documented in the PR:

• engine_daemon_image_actions_seconds (Histogram) — "The number of seconds it takes to process each image action", labels: delete, push, history, pull
• engine_daemon_health_check_start_duration_seconds (Histogram) — "The number of seconds it takes to prepare to run health checks"
• swarm_node_info (LabeledGauge) — "Information related to the swarm", labels: swarm_id, node_id