n8n-guard
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| N8N_URL | No | Base URL of the n8n instance. | http://localhost:5678 |
| N8N_API_KEY | No | Public API key (Settings, n8n API). Needed for the execution and workflow checks. | |
| N8N_SQLITE_PATH | No | Path to database.sqlite, needed for datastore checks on SQLite. | |
| N8N_POSTGRES_URL | No | Postgres connection string for datastore checks; takes precedence over N8N_SQLITE_PATH. | |
| N8N_GIT_REPO_PATH | No | Checkout holding exported workflow JSON, needed for the drift check. | |
| N8N_GUARD_TIMEOUT_MS | No | Per-request API timeout in milliseconds. | 15000 |
| N8N_GUARD_DB_WARN_BYTES | No | Datastore warning threshold in bytes (default 1 GB). | 1073741824 |
| N8N_GUARD_STUCK_MINUTES | No | Minutes in `running` before a run counts as stuck. | 15 |
| N8N_GUARD_DB_CRITICAL_BYTES | No | Datastore critical threshold in bytes (default 4 GB). | 4294967296 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| instance_infoA | Reachability, configuration and workflow inventory of the monitored n8n instance. Start here: it shows which checks are configured and which are not. |
| db_healthA | Size, per-table breakdown and growth rate of the n8n datastore, plus an estimate of when it hits the critical threshold. Works on SQLite and Postgres. The n8n Public API has no endpoint for this. |
| stuck_executionsA | Executions sitting in |
| retention_reportA | How many execution records are stored, how far back they go, and what pruning to a retention window would free. Read-only: n8n-guard never deletes anything. |
| git_driftA | Diffs the workflows in the running instance against the JSON exports in a git checkout: live but unexported, exported but deleted, and content that has drifted apart. |
| guard_runB | Runs every check in one resilient sweep. Individual steps retry and, if they still fail, are reported as findings while the remaining checks finish. Use this for scheduled monitoring. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 6 tools
Each check targets a distinct n8n failure mode: instance inventory, datastore growth, stuck executions, retention, and git drift. guard_run is explicitly an aggregate sweep, so its overlap with the individual checks is intentional and clearly described.
All tool names use lower_snake_case and concise check-oriented identifiers. guard_run is slightly more action-oriented than the others, but the naming convention remains consistent throughout.
Six tools is well-scoped for a read-only n8n monitoring server. Each tool maps to a meaningful check or aggregate runner, with no filler.
The surface covers the core monitoring lifecycle: instance discovery, DB health, stuck executions, retention, git drift, and a resilient aggregate run. Minor gaps remain around explicit alerting/escalation or worker/queue health, but the main guard workflows are covered.