Skip to main content
Glama
mcp-oncall-retainer

mcp-oncall-retainer

mcp-oncall-retainer

MCP server for the Agent Incident On-Call Retainer - the staffed escalation layer. When an SLA-breaching event fires, a named responder owns triage, mitigation, and post-mortem. Distinct from the SLA contract service: this is the people.

Pricing in the V3 catalogue: $49-199/mo base; $199-500/mo guaranteed-response tiers; desks indicated willingness above $400 when bundled with SLA terms. The "can a small operator truly staff 24/7?" objection is answered with published staffing transparency, follow-the-sun coverage and credits.

Tools

get_oncall_status

Resolve who is on call at a given moment from the follow-the-sun rotation schedule.

Parameter

Type

Default

Description

team

string enum core/defi

core

On-call team to query

at

string

now

ISO-8601 timestamp to resolve on-call for

Returns the active responder, shift window, minutes into/remaining in shift, the handoff partner and its handoff time, the next rotation, the team's escalation ladder, and staffing-transparency metrics (rotations per day, distinct responders, follow-the-sun coverage hours, whether coverage is 24x7, single-operator risk).

create_incident

Raise an incident and persist it.

Parameter

Type

Default

Description

title

string

required

Incident title

severity

string enum sev1/sev2/sev3

sev2

Severity

service

string

unspecified

Affected service or agent

summary

string

-

What happened

team

string enum core/defi

core

Owning on-call team

raised_at

string

now

ISO-8601 raise time

acknowledged_at

string

-

Optional acknowledgement time

mitigated_at

string

-

Optional mitigation time

resolved_at

string

-

Optional resolution time

status

string enum open/acknowledged/mitigated/resolved/postmortem

open

Initial status

The responder is resolved from the rotation at the raise time. Acknowledge and mitigate deadlines are derived from the severity multipliers and the team's response tier; the response tier's mitigation window doubles as the SLA recovery window, and any breach is linked to the Agent Failure-Recovery SLA service for credit computation. Timestamps that precede raised_at are rejected with status: "error".

get_incident_history

Incident history, newest first.

Parameter

Type

Default

Description

team

string enum

-

Filter by team

severity

string enum

-

Filter by severity

service

string

-

Filter by affected service

status

string enum

-

Filter by status

since

string

-

Only incidents raised at or after this ISO timestamp

limit

integer

50

Maximum entries (1-500)

Each incident carries its triage timeline (raised / acknowledged / mitigated / resolved with minutes-from-raise) and its post-mortem link.

get_response_metrics

Response metrics computed from the incident record set.

Parameter

Type

Default

Description

window_days

integer

30

Look-back window in days (1-365)

team

string enum

-

Filter by team

service

string

-

Filter by service

Returns mean/median/p95/min/max time-to-acknowledge and time-to-mitigate, acknowledgement and mitigation SLA compliance percentages, escalation rate, incident volume by severity, and a daily time series. All figures are computed from the actual incident records with Python statistics.

Related MCP server: StackWitness

Installation

pip install -r requirements.txt

Usage

python main.py

State is persisted as JSON in ./state/oncall_state.json. Override the location with MCP_ONCALL_RETAINER_STATE_DIR.

Connecting to an MCP client

{
  "mcpServers": {
    "oncall-retainer": {
      "command": "python3",
      "args": ["/tmp/v3-services/mcp-oncall-retainer/main.py"]
    }
  }
}

License

MIT

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables PagerDuty incident response operations including listing incidents, acknowledging and resolving incidents, looking up on-call schedules, and listing services.
      MIT
    • F
      license
      Not graded
      quality
      D
      maintenance
      Let agents read and manage your org’s independent monitors, incidents, status pages, and alerts (honest measured state, no false greens).
      -
    • A
      license
      A
      quality
      B
      maintenance
      Enables interaction with Jira Service Management Operations alerts and on-call schedules, including searching alerts, reading notes and activity logs, acknowledging, closing, annotating, adding responders, and querying current and next on-call rotations.
      28
      81 npm
      4
      Apache 2.0