6.15 SLAMonitor

1. Bot Identity

2. Purpose

SLAMonitor tracks service-level objectives committed to internally and to users, measures error-budget burn rate, and emits alerts when burn rate approaches the SLO budget limit. Retained 7 years as a compliance-grade availability record.

3. Why This Bot Matters

• No SLO tracking

  Availability and latency regressions go undetected until users complain; SLA breach evidence is unavailable for compliance.

• Error budget burn not tracked

  The team consumes the entire error budget without realising it; no time left for planned maintenance.

No worked examples on this bot yet. Worked examples are optional but strongly recommended — they turn an abstract failure mode into something a developer can verify in a fixture.

4. Required Polymarket Inputs

5. Required Internal Inputs

6. Parameter Guide

7. Detailed Parameter Instructions

{ "slo_definitions": {"fill_latency_ms_p99": 500, "fill_success_rate_pct": 99.5, "uptime_pct": 99.9} }

if metric_value > slo.target: budgetConsumer.record(slo.name)

{ "burn_rate_alert_pct": 5.0 }

if hourly_burn > p.burn_rate_alert_pct: emit('SLO_BURN_RATE_EXCEEDED')

8. Default Configuration

{
  "bot_id": "gov.slamonitor",
  "version": "0.1.0",
  "mode": "shadow_only",
  "defaults": {
    "slo_definitions": {
      "fill_latency_ms_p99": 500,
      "fill_success_rate_pct": 99.5,
      "uptime_pct": 99.9
    },
    "burn_rate_alert_pct": 5.0,
    "publish_to_user": true,
    "auto_freeze_on_breach": false
  }
}

9. Implementation Flow

1. Scrape Prometheus metrics from all fleet bots every 60 seconds.
2. For each SLO definition, compute current compliance and error-budget consumption.
3. Compute hourly burn rate as (errors_in_last_hour / monthly_budget * 100).
4. If burn_rate > burn_rate_alert_pct, emit SLO_BURN_RATE_EXCEEDED alert.
5. If error budget is exhausted, emit SLO_BREACH_DETECTED and optionally freeze deployments.
6. Emit SettlementReport(event_type=SLO_STATUS) every hour with all SLO compliance metrics.
7. Retain SettlementReport records for 7 years as compliance-grade availability evidence.

10. Reference Implementation

Pseudocode is language-agnostic. FETCH = read input. EMIT = produce output. IF/THEN/ELSE = decision. Translate directly to TypeScript, Python, Go, or Rust.

// ---- SCRAPE LOOP (every 60s) ----
FUNCTION scrapeAndEvaluate():
  metrics = FETCH internal.metricsStore.GET({bots: 'all', window: '1m'})
  IF metrics IS NULL:
    EMIT SettlementReport(event_type='SLO_DATA_GAP')
    alerting.emit('SLO_METRICS_UNAVAILABLE')
    RETURN

  compliance = {}
  FOR slo IN config.slo_definitions:
    actual = metrics.get(slo.name)
    compliant = (actual <= slo.target) IF slo.type == 'max' ELSE (actual >= slo.target)
    compliance[slo.name] = {target: slo.target, actual: actual, compliant: compliant}
    IF NOT compliant:
      errorBudget.record(slo.name, violation=True)

// ---- HOURLY REPORT ----
FUNCTION emitHourlyReport(windowStart, windowEnd):
  burnRate = errorBudget.hourlyBurnRate()
  IF burnRate > config.burn_rate_alert_pct:
    alerting.emit('SLO_BURN_RATE_EXCEEDED', {burn_rate: burnRate})
  IF errorBudget.exhausted():
    alerting.emit('SLO_BREACH_DETECTED')
    IF config.auto_freeze_on_breach:
      deploymentManager.freeze()
  EMIT SettlementReport(event_type='SLO_STATUS',
    window_start=windowStart, window_end=windowEnd,
    slo_compliance=compliance,
    error_budget_consumed_pct=errorBudget.consumedPct(),
    hourly_burn_rate_pct=burnRate,
    retained_until=now() + days(2555))

SDK calls used

• internal.metricsStore.GET({bots, window})
• errorBudget.hourlyBurnRate()
• alerting.emit('SLO_BURN_RATE_EXCEEDED', metadata)

Complexity: O(S) per scrape cycle where S = SLO count; O(1) for hourly report

11. Wire Examples

Input — what arrives on the wire

{
  "label": "Prometheus metrics scrape",
  "source": "internal.metrics_store",
  "payload": {
    "fill_latency_ms_p99": 312,
    "fill_success_rate_pct": 99.8,
    "uptime_pct": 100.0,
    "scraped_at_ms": 1746792060000
  }
}

Output — what the bot emits

{
  "label": "SettlementReport — SLO_STATUS",
  "payload": {
    "report_id": "stl_sla_01HX9Z",
    "event_type": "SLO_STATUS",
    "error_budget_consumed_pct": 1.2,
    "hourly_burn_rate_pct": 0.8,
    "report_kind": "SettlementReport",
    "topic": "polytraders.reports.settlement",
    "retained_until": "2033-05-09"
  }
}

12. Decision Logic

13. Standard Decision Output

This bot returns a SettlementReport object. See SettlementReport schema.

{
  "report_id": "stl_slamonitor_01HX9Z",
  "bot_id": "gov.slamonitor",
  "event_type": "SLO_STATUS",
  "window_start": "2026-05-09T09:00:00Z",
  "window_end": "2026-05-09T10:00:00Z",
  "slo_compliance": {
    "fill_latency_ms_p99": {
      "target": 500,
      "actual": 312,
      "compliant": true
    },
    "fill_success_rate_pct": {
      "target": 99.5,
      "actual": 99.8,
      "compliant": true
    },
    "uptime_pct": {
      "target": 99.9,
      "actual": 100.0,
      "compliant": true
    }
  },
  "error_budget_consumed_pct": 1.2,
  "hourly_burn_rate_pct": 0.8,
  "report_kind": "SettlementReport",
  "topic": "polytraders.reports.settlement",
  "retained_until": "2033-05-09"
}

14. Reason Codes

15. Metrics & Logs

Metrics emitted

Alerts

16. Developer Reporting

{
  "bot_id": "gov.slamonitor",
  "event_type": "METRICS_SCRAPED",
  "slo_name": "fill_latency_ms_p99",
  "actual_value": 312,
  "target_value": 500,
  "compliant": true,
  "scraped_at_ms": 1746792060000
}

17. Plain-English Reporting

18. Failure-Mode Block

19. Failure-Injection Recipes

20. State & Persistence

Cold-start recovery

On restart, reload error budget state from last committed SettlementReport.

21. Concurrency & Idempotency

22. Dependencies

Depends on (must run first)

Emits to (downstream consumers)

Sibling bots (same OrderIntent)

External services

23. Security Surfaces

Abuse vectors considered

• Manipulating metrics to suppress SLO breach detection

Mitigations

• Metrics store is read-only for SLAMonitor; write access is restricted to fleet bots only

24. Polymarket V2 Compatibility

API surfaces declared

Networks supported

25. Versioning & Migration

Migration history

26. Acceptance Tests

Unit Tests

Integration Tests

Property Tests

27. Operational Runbook

SLAMonitor incidents require rapid response when error budget burns faster than planned. P1 if budget is exhausted; P2 for high burn rate.

On-call actions

Manual overrides

• —

Healthcheck

/internal/health/slamonitor → green if Metrics store reachable; all SLOs compliant; burn rate < burn_rate_alert_pct; red if Metrics store unreachable or any SLO budget exhausted

28. Promotion Gates

A bot does not advance to the next readiness state until every gate below is green. Gates are observable from production data — no subjective sign-off.

Promote to Shadow

Promote to Limited live

Promote to General live

29. Developer Checklist

Ready-to-ship score: 27/27 sections complete · 100%