openstack-incidents
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AGENT_DATABASE_URL | No | Database connection string for PostgreSQL. Default is postgresql://incident:incident_dev_password@127.0.0.1:55432/incident_agent. |
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": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_instance_summaryC | Return signals only for the specified instance/request pair. |
| search_logsC | Read at most 100 matching logs. This tool never exposes evaluation labels. |
| get_timelineC | Return bounded opening and terminal events of the specified operation. |
| search_runbooksC | Find citeable Runbook chunks; guidance is not incident evidence. |
| get_case_memoryB | Read approved drafts for this instance/request pair as historical context. |
| validate_evidenceC | Verify citations match both the investigated instance and request. |
| save_ticket_draftA | Persist a draft in PostgreSQL only when the caller explicitly passes approved=true. |
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 7 tools
Most tools target distinct resources/actions (validate_evidence, save_ticket_draft, search_logs, search_runbooks, get_timeline), but the read-side tools get_instance_summary, get_timeline, search_logs, and get_case_memory all return signals/context for the same instance/request pair and could be confused without careful reading. The descriptions disambiguate well enough (e.g. timeline = opening/terminal events, case memory = approved drafts).
All names follow a consistent snake_case verb_noun pattern (get_*, search_*, validate_*, save_*), so the operation type is predictable at a glance. No mixed conventions or vague verbs.
Seven tools is well-scoped for an incident investigation workflow, covering read/evidence/draft concerns without redundancy. Each tool earns its place.
The surface covers the core investigation lifecycle: gather signals, timeline, logs, runbooks, validate evidence, persist a draft, and recall prior drafts. Minor gaps exist (no tool to fetch/close the ticket itself or delete/update drafts), but agents can work around them.